o

ft *</>#
Communication
3^ *
w = S kills-I

INTRODUCTION
Communication simply means exchange of information through signs, signals,
speaking, writing or using some other medium. The word communication is derived
from the Latin word “ communicare” , which means to share. Communication
involves at least tw o people— a sender and a receiver. Communication is
successful only when there is a common understanding between the sender and
receiver.
Communication is something we all do on a daily basis without paying much
attention. It started hundreds of years ago and has evolved into writing, listening,
speaking and reading as four key components.
Elements of communication cycle:
(i) sender (ii) ideas (iii) encoding
(iv) communication channel (v) receiver (vi) decoding
(vii) feedback

THE COMMUNICATION CYCLE

It consists of six key concepts:
1. Sender: The person who starts
the communication process by
Sender — *■ Encoding — *■ M essage Decoding - Receiver
choosing a message to send,
someone to send the message
to and a channel/medium to
Fig. 1.1 : Communication Cycle
send the message.
2. Encoder: The sender is also the encoder who transforms a thought or an idea
into words or body language to communicate to the receiver.

C o m m unicatio n S kills - I ES.1

3. Channel: The medium through which information from the sender and
encoder goes through to the decoder and the receiver. It could be oral, written
or visual.
4. Noise: There could be potential interruption to a message on the way from the
sender to the receiver. This is known as Noise.
• Internal noise: When the sender uses a different tempo or slangs and the
receiver is not able to understand the words.
• External noise: When the noise is due to an external or uncontrollable reason
such as talking in a crowded place.
5. Decoder: The content of the message being understood by the person at the
other end of the communication.
6. Receiver: The person who gets the sender’s message.
Another concept (‘feedback’) was later added.
Feedback: When the receiver responds to the sender to give feedback on whether
the message sent by the sender was understood in its entirety and intent.

IMPORTANCE OF COMMUNICATION SKILLS

Developing effective communication is a skill which will serve you well in life. Here
are some of the key benefits of good communication skills:
• Managing problems and conflicts: You are able to better manage issues
with family members, friends and colleagues if you have good communication
skills.
• Better relationships: Listening carefully and providing feedback helps maintain
transparency and improves relationships.
• Higher productivity: When people around you understand your expectations
and how they need to work with you, productivity increases automatically.
• Less stress: Since there will be less misunderstandings, it will create less stress
for you in life.
• Respect and confidence: If you are a good communicator, more opportunities
will open up for you in life and you will gain respect of everyone around
you.

ES.2 E ssentials of A rtificial Intelligence - I X

Types of Communication
Communication can be any form of signal or message. It could be in any format. Let
us understand it through a simple example: When a baby cries, it sends a message
to the parents that it needs attention. This is a form of non-verbal communication.
When we start to speak, it is known as verbal communication.

When we start going to school, we start doing w ritten communication. Even when
we enter our professional careers, written and verbal communication remain the
most important aspects of our daily lives.

Advantages of Non-verbal Communication

1. Supports the main message: Non-verbal cues can add weight and supplement
the main message or idea.
2. Visual appeal: Information can be visualized in a simple manner which is easily
understood.
3. Substitute for verbal communication: In case of barriers such as noise or distance,
we can use gestures to communicate such as nodding instead of a yes.
4. Faster communication: Signs and gestures can communicate much faster than
verbal communication.

Disadvantages of Non-verbal Communication

There are some limitations too of non-verbal communication:
1. Non-specificity: Non-verbal communication can sometimes be broad in its
message and perceived differently by different people. Its meaning varies in
various cultures and demographics.
2. Continuity: It is difficult to carry out a long conversation through only non-verbal
communication.
3. Distracting: It may be distracting to view multiple gestures and understand
them at the same time. In case someone is communicating non-verbally
as well as through verbal communication, it gets a little complicated.

C ommunication S kills —I ES.3

4. Cultural variations: Gestures can mean different things in different countries.
For example, waving in the air may be considered friendly in one culture while
rude in another.

Pros and Cons of Written Communication

Pros
• A proper record and trail of the communication
• Saves resources by eliminating the need to meet physically
• Presentable means of communication with no ambiguity
• Gives the receiver sufficient time to think, act and react

Cons
• Consumes time in preparation and documentation
• Non-confidentiality - anyone who has access to the document can read it unless
it is encoded
• No scope of clarification and is subject to possible misunderstanding
• Not useful in emergency situations

Pros and Cons of Verbal Communication

Pros
• Faster mode of communication with real-time feedback
• Conversations can be recorded and saved for future use and as evidence
• Effective for persuasion and convincing
• Helps in making better relationships

Cons
• Barrier to language and translation difficulties
• Difficult to convey long messages
• No option to recall or edit the message
• Subject to cultural misinterpretation

ES.4 E ssentials of A rtificial I ntelligence - I X

 Role-play a communication process in a Hospital
Let us say you are handling the front office desk at a hospital. Imagine that a
patient enters and wants to book an appointment to meet a doctor. Let us look at
the conversation flow:
Patient: Good morning!
Front Office Desk: Hi, Good morning!
Patient: I am looking to meet Dr. Priyanka.
Front Office Desk: Sorry sir, she is not here at the moment.
Patient: May I know when will she be here?
Front Office Desk: Let me try to call her and update you. You can wait in the lobby please.
Patient: I am in a bit of a hurry.
Front Office Desk: Let me see what I can do.
Can you identify the components of communication here and improve on the
communication from both the sender’s and receiver’s perspective?
Write your version below:

BODY LANGUAGE
Body language is an important aspect of non-verbal communication where physical
behaviour is used to convey a message or information.
Body language includes:
• Facial expressions • Posture
• Hand gestures • Touch
• Use of space • Eye movement

It is also known as “ kinesics” /kinaesthetics.

C ommunication S kills - I ES.5

 BODY LANGUAGE
Nodding, to indicate
yes, is not universal.

0
Grinning or smiling
in some cultures
could be interpreted
as being pushy.
Don't point with
your finger, it may
e considered rude.

Don't do the OK sign; the Be careful where you

gesture means different point your feet, and
things in other cultures. avoid pointing your toes
or soles at another person.

Do’s and Don’ts of Non-verbal Messages and Body Language

D o’s
• Make eye contact but don’t stare
• Have a firm handshake but no hand-crushing
• Have a calm behaviour
• Be natural with your gestures and posture
• Smile and act natural
• Listen intently

Don'ts
• Avoid touching your face
• No leaning back and sitting casually
• Do not sit in cross-armed position
• Do not avoid making eye contact
• Hand movements/gestures are a no-no

ES.6 E ssentials of A rtificial Intelligence - I X

 YOU KNOW?

VERBAL 35%

NON-VERBAL 65%
Facial Expressions
Tone of Voice
M ovem ent
A ppearance
Eye Contact
Gestures
Posture ^

Perspectives in Communication
Perspectives are various ideas or our ways of thinking. They can affect our
communication. For example, if you really like someone, you will most likely
ignore their silly acts but if you don’t like someone, you may exaggerate their
every little fault and think that they are doing something wrong. It is basic
human nature.

Factors affecting Perspectives in Communication

Factor How the factor can become a barrier
Language Language is the basic factor in communication as it forms the basis of all
communication. Imagine if two people from different nationalities interact
without knowing each other’s language, how difficult it would be. It affects
all forms of communication like verbal and written.
Visual perception Visual perception is the ability to see, organize and interpret one’s environment.
It also includes completing partially drawn pictures with visual perception.
Past experience Many a time we draw on our past experiences to communicate in the
present as well. For example, I didn’t score well in a certain subject so this
time too I may not do well.
Prejudice Fixed notions that you need to study from certain sources and questions in
the exam will come from there only is a wrong notion and can be misleading.

Communication S kills- I ES.7

Feelings A person’s feelings and emotions can sometimes lead to wrong
communication, e.g., when we don’t feel well, we don’t talk properly or say
things we don’t actually mean.
Environment Environmental distractions may also lead to difficulty in understanding. For
example, if there is noise on the street and someone is talking to you on
phone, it is difficult to hear the other person properly.
Personal factors Various personal habits may lead to difficulty in communication. Feelings
of anxiety, depression and anger may affect proper communication.
Culture Different cultures have different meanings of signs and gestures such as
showing a thumb may mean “good job” done for some people but it may be
considered insulting in another culture.
As w e have seen above, it is very im p o rta n t to have a p p ro p ria te techniques o f
com m un icatio n to m ake sure th e message is delivered in th e fo rm and in te n t
desired by th e sender.

Factors affecting communication at the workplace

Communication at the workplace is critical for smooth flow of work. There are
few factors influencing communication at the workplace:
• Cultural diversity
• Misunderstanding of message
• Emotional difference
• Past experiences in life
• Educational/intellectual difference
• Group and cultural dynamics
• Positional/hierarchical difference among the personnel
• Functional relationship between the sender and receiver

WRITING SKILLS
Parts of a Sentence
A complete sentence consists of two parts: a subject and a predicate. The subject is
what the sentence is about, while the predicate tells us something about the subject.

Subject
X
The manager has changed all of his staff recently
t
Predicate

ES.8 E ssentials of A rtificial Intelligence - I X

Types of Sentences
There are four basic types of sentences, depending on the purpose of the speaker
or writer.
1. Look at the red roses! This is an Imperative sentence which gives a command.
2. The rose is red. This is a Declarative sentence which states a fact, opinion or idea.
3. Is the rose red? This is an Interrogative sentence which asks a question.
4. I love the red rose! This is an Exclamatory sentence which expresses a strong
emotion.

Imperative r ends with a full stop (.) or an exclamation (!)

Declarative ends with a full stop (.)

Interrogative 9 ends with a question mark (?)

Exclamatory 9 ends with an exclamation (!)

A sentence is a sequence of words which conveys a complete thought.

Can you add the right punctuation marks to create the sentence category on the right.
Types of sentences-Recording Sheet

Add Punctuation Challenge (What is the sentence type?)

1. Pass the book to the student imperative
2. There is a class upstairs exclamatory
3. What is Kartik doing tomorrow afternoon interrogative
4. Finish your homework before yougo to play imperative
5. He is very intelligent declarative
6. Attention exclamatory
7. He is wearing a coat declarative

C ommunication S kills—I ES.9

Simple, Compound, Complex
1. A simple sentence containing a subject and a verb. It expresses a complete idea.
subject verb object
simple sentence ^K artik hvorks at DcodeAI.

2. A compound sentence contains two independent clauses joined by a conjunction.

The first part of the sentence and the second part make sense independently.
^independent clause coordinator ^independent clause
compound sentence 1K a rtik works a t D codeA I and 'K a rtik leads th e technology division.

3. A complex sentence uses an independent clause combined with one or more

dependent clauses. The dependent clause cannot stand alone as a sentence.
dependent clause independent clause

he still needs to do household work Even though Kartik works at DcodeAI

complex sentence
Even though Kartik works at DcodeAI, he still needs to do household work.

PARTS OF SPEECH
Parts of Speech Function
Nouns Naming people, places, things and ideas/concepts
Adjectives Describing nouns and pronouns
Adverbs Modifying verbs, adjectives and adverbs
Verbs Expressing a physical action or state
Prepositions Showing relationships of words and phrases
Pronouns Replacing nouns
Conjunctions Joining words, phrases and clauses
Interjections Showing strong emotions or feelings

A noun names a person, place, idea or thing.

An adjective describes a noun, adds some bling.
A verb shows action, experience or state of being.
An adverb modifies a verb, a new way of seeing.
The conjunction joins two or more words or thoughts.
But to show a relationship a preposition must be brought.
The pronoun is a replacement for the good old noun.
Interjection to express emotion—feeling up or down.

ES.10 E ssentials of A rtificial Intelligence - I X

ARTICLES
In English, there are three articles— a, an and the.

ARTICLES : A/AN/THE
A/AN THE
w>
• Used for objects that are not specific or • Used for specific objects or objects that
one of several things of a similar type both the speaker and listener know
e.g., I need to buy a new phone. e.g., Please post the letter today.

• Used the first time we introduce an

object • Used when we mention the object again
e.g., Let’s try a new recipe. e.g., The recipe had precise instructions.

Parts of Speech
Article Adjective Noun Verb Preposition
* ^ v '
>• ' ✓ - \
The old woman was sitting at
Article Noun Conjunction Adverb
/ •
✓ *x / /
' /
the railway station patiently
. .

Verb Preposition Pronoun Noun Verb

V ' / / / \

\ ' /
waiting for ler train to arrive.

PHRASES
A phrase is a group of two or more words that work together to convey a meaning.
However, a phrase is not a complete sentence.
Writing Effective Paragraphs

Sentences come together to form a paragraph. A paragraph is

PARAGRAPH WRITING
held together by a ‘topic sentence\ A topic sentence tells readers
Topic sentence
what the particular paragraph is about. All sentences which come Supporting ideas
Adding details
after it help in extending the idea or topic and provide more Concluding sentence

facts and information. They are known as ‘supporting ideas'.

Then we keep adding details to the paragraph and finally a concluding sentence
is added.

_____________

Can you identify the four categories of sentences from the below paragraph:
Open Government Data Platform data.gov.in has a very robust policy of reviewing its content through
routine testing of the web pages and the content. Since this is a platform which boasts of data sets and
catalogue, the Project Management Unit (PMU) is frequently in the process of updating and reviewing the
data sets and catalogues in order to provide up-to-date data sets to its users. The feedback and engagement
policy with the users itself provides an effective measure of reviewing and checking the content on the
platform. Monitoring of platform is done on a regular basis. It is important for anyone using open data
to refer to this platform as it is the most authentic source of data available in the country. Do you know
it has more than 50,000 resources!
Whatever data requirements you may have for your next Al project, it can help you. Can you think of a
problem for which you were looking for a data set?

^The Last Straw

This is a creative writing exercise where a given instruction needs to
be changed to a funny statement that is deemed more effective. It is a
good exercise for w riting short instructions/signs to be used in public
places. You will see how humour adds flavour to the conversation.
Activity Steps:
• Refer to the example image.
• Research popular statements used in public places or think of one you m ight have
come across that is suitable for this exercise. Consider instructions that command
or dictate something.
• Convert the original statement into something funny to increase engagement.
• Consider two instructions and make two statements like the given example.
• You can use the internet for research or inspiration.
• Allocate yourself 20 minutes for research and to complete this exercise.
• Discuss with your classmates and vote for the best statement.
v______________ :___________________________________________________________
(Source-. CBSE)

ES.12 E ssentials of A rtificial Intelligence - I X

 What do you think of other students’ statements? Which one was your
favourite and why? What strategy did you follow to make the instruction
more engaging? Were you inspired by examples of others? Would you
consider changing your statement to improve it?

Objective Type Questions I

1. Which of these is a series of sentences that are organized, coherent and related to
one single topic?
(a) Essay (b) Sentence (c) Verb (d) Paragraph
2. What is known as the response of the receiver to the message?
(a) Encoding (b) Feedback (c) Decoding (d) Channel
3. Which of these is not an example of body language?
(a) Facial expressions (b) Posture
(c) Gestures (d) Verbal communication
4. Which of these is not a barrier to communication?
(a) Personal barriers (b) Language barriers
(c) Gender barriers (d) Financial barriers
5. Which of these is not a factor affecting communication at workplace?
(a) Group affiliations (b) Cultural diversity
(c) Misunderstanding (d) Financial differences
6. Which of these is not a type of sentence?
(a) Declarative (b) Imperative
(c) Normal (d) Interrogative
7. What form of communication is transmission of information and ideas using symbols
and imagery?
(a) Verbal (b) Visual
(c) Written (d) None of these
8. Which of these is the oldest form of communication?
(a) Written (b) Oral
(c) Email (d) WhatsApp
9. The process of communication involves a .......................... transmitting a message
through a ..........................to th e .............................
10. Communication is successful only when both sender and receiver have a common
understanding. (True/False)

C ommunication S kills - I ES.13

: Subjective Type Questions •

1. What do you understand by communication and why is it important in our life?

2. What are the pros and cons of written, verbal and non-verbal communication?
3. What is the difference between verbal and non-verbal communication? Give examples.
4. Write five Do's and Don'ts of effective body language.
5. What are the factors affecting our perspectives in communication?
6. Explain the various parts of speech. Use them in a paragraph and show each
component.
7. What are the four types of sentences? Explain with one example for each.
8. What is the importance of feedback in the communication process? Explain various
sources of collecting feedback.
9. Draw a diagram of communication cycle and list down the importance of
communication at the workplace.
10. Have a discussion with your parents and teacher and note down their experiences on
factors affecting communication at the workplace and how they handle it. List down
some good practices to be followed in workplace communication.

ES.14 E ssentials of A rtificial I ntelligence - I X

 o
# * < /> '*

S elf-Management
If 3 S kills-I

INTRODUCTION
Self-management refers to management of or by oneself. It means taking
responsibility for one’s own behaviour and wellness. It also means to manage people
around you and keeping healthy relationships with one’s own self and others around
us. It extends beyond family and looks at the greater good of society. A self-managed
person manages his/her life in a meaningful manner and supports the community.
Self-management is important because:
• All individuals should be self-sufficient and independent.
• It instils confidence in an individual that they can solve their problems on their
own.
• It helps in creating a framework for managing one’s life in an organized manner
without undue stress.
Self-management is a combination of our habits and behaviour.
Some of the good habits include:
• Maintaining body hygiene
• Eating a balanced diet
• Regular exercise
• Proper sleep and rest
• Meditation for mind management
Some of the good behaviour traits include being:
• Kind and friendly
• Respectful to others
• Considerate of others’ feelings
• Polite in conversations
• Of a caring and sharing nature
S elf -M anagement S kills - I ES.15
 Fig. 2.1 : Exhibitors of Self-Management

POSITIVE RESULTS OF SELF-MANAGEMENT

As a student, you may be faced with a lot of stress due to homework, class schedule
and peer pressure. You may not always feel comfortable talking to parents or
teachers for guidance. In these situations, self-management techniques are helpful
to manage stress and tackle your problems without anyone around to help you.

/ \

Once you start using self-management techniques, you will realize the following
benefits:
• Managing tasks independently
• Ownership and accountability leading to self-confidence
• Self-monitoring and discipline
Self-management helps in improving our goal-setting and finding out ways to
achieve them. It also helps in defining metrics for achieving a goal and changing
methods if the desired results are not achieved. For example, if you are targeting
100% marks in Mathematics, you will define steps that you need to undertake to
reach this goal from your current skill level with defined milestones. In case your
approach is not working and you are not reaching 100%in defined timelines, you
will relook at your approach and improve upon it.

ES.16 E ssentials of A rtificial Intelligence - I X

Self-management helps in reinforcing positive behaviour and creates a healthy
environment for everyone around you. A self-managed person can make a big
impact on society.

SELF-MANAGEMENT SKILLS
Following are the self-management skills that individuals need to develop:
Self-confidence: It means to trust one's capabilities and skills to achieve one's
goals. It means to not doubt in one’s abilities to lead a successful life. People with
self-confidence inspire others around them.
Starting a new initiative or working on an innovative project can build your confidence.
Stress Management: It means to manage any unforeseen situations or circumstances
which you cannot control directly with your actions. As a student, you may
experience stress because of change in class timetable or surprise tests. There
may be stress due to peer pressure and parental expectations.
Managing your time schedule and handling expectations through meditation and
communicating with your parents may help you in easing stress.
Independent Working: This refers to the skill of being able to work on one’s own
without constant guidance and supervision. This leads to confidence in one’s ability
to do tasks independently.
Preparing for a test, a competition or learning something new on your own shall
help enhance self-reliance.
Team Player: It is a person who actively contributes to their team to achieve a
common goal. While working on a common project, you develop interpersonal
and group management skills as well.
Time Management: It is the task of managing time at hand to increase productivity,
efficiency and effectiveness.
As a student, you need to juggle between studies, hobbies and extracurricular
activities. Time management becomes an important skill to keep a balance in life.
Self-Motivation: Motivation is defined as the key force which drives us to achieve
our goals. As a student, you need to keep yourself motivated to achieve good
scores and have a health balance in life.
Personality Management: Dressing properly, maintaining personal hygiene and
being a good conversationist create a positive impression on others. Your personality
is the first thing people observe when they see you.

S elf -M anagement S kills - I ES.17

Self-confidence: Self-confidence is the ability to have trust on one's abilities and
skills.
Factors that help in building self-confidence:
• Social: Interaction with family members, teachers and peers help in building
meaningful relationships which leads to confidence-building.
• Cultural: Cultural factors include family values, moral values and belief system.
Adhering to a system of values helps in giving comfort and developing confidence.
• Physical: Physical self-efficacy and physical activity can lead to individuals
thinking better about themselves. People who exercise and take good care of
their body feel happy and contented.

How to Build Self-Confidence

Here are a few tips to help you build self-confidence:
• Stay positive: Think positive and get rid of any negative feelings. Whenever you
feel overwhelmed with problems, take deep breaths and think of positive things.
• Have gratitude: Be thankful and grateful to everyone around you and always
be kind to people.
• Personal hygiene: Take a bath every day and keep your hair, teeth, fingers and
skin clean to take care of your body and mind.
• Spend quality time w ith loved ones: Interact with people who care for you and
spread positivity in life. Spend time with those who propel you forward in life
and support in achieving your goals.

STRESS STRESS
CAUSAL
Stress refers to a “ physical, mental or emotional AGENTS
strain or tension” .
Stress is one of the key reasons for decline in
physical and mental wellness. A little amount
^Social^- J Financial
Fi

of stress is needed in life but excessive

worrying can lead to disorders.
IphysicalL-^ Mental j
Stress Management
Stress management refers to focusing on maintaining a healthy body and mind to
handle stressful situations. Stress is an unavoidable part of life, be it due to a family
issue, a fight with a friend or an approaching exam.

ES.18 E ssentials of A rtificial I ntelligence - I X

Let us identify the sources of stress:

Importance of Stress Management

As can be seen, we are surrounded by various sources of stress. It is important to
be able to cope with stress to lead a healthy and happier life to ward o ff illnesses
later in life. Stress management can lead to the following benefits:
• Improves mood and general well-being
• Boosts immune system
• Helps in increasing lifespan
• Checks distractions in life
• Prevents psychological and other behavioural disorders

Stress Management Techniques

Stress is a mental state wherein an individual may feel overwhelmed with a situation.
It affects different people differently. Some of the symptoms of stress may be
common while others may vary from person to person.
Some physical signs which may point to stress are:
• Breathlessness • Dry mouth or choking feeling
• Indigestion • Nausea
• Fatigue • Sweaty palms
• Cold hands and feet
At the mental level, the symptoms may include:
• Irritation • Loneliness
• Mood swings • Anxiety
• Boredom

S elf- M anagement S kills —1 ES.19

Self-awareness about symptoms of stress can help in managing it better. It is
important to speak to well-wishers like family and friends, if required. Apart from
sharing and thinking positively, engaging in the following activities can also help
manage stress:
• Physical exercise: Physical exercise in the form of walking, running or going to
the gym can be helpful in relieving stress as it helps stabilize mood, improves
self-esteem and induces sleep after a long day at work.
• Yoga: Yoga includes a series of asanas, postures and breathing exercises to gain
control of your body and mind. There are various types of yoga which help calm
both the body and the mind.
• Meditation: It is a practice where an individual uses a set of techniques and
focuses on a particular object or thought to achieve a calm state of mind, thus
reducing stress.
• Recreational activities: Engaging in activities like watching movies, playing games,
singing, dancing or arts can help one get to a calmer state of mind and be cheerful.
• Going on vacations w ith family and friends: Vacations with your loved ones
provide an opportunity to appreciate different flavours of life and take a
break from the mundane daily routine. They are very useful in de-stressing and
rejuvenating oneself.
Taking nature walks: Taking nature walks in the woods or in the countryside
relaxes the brain and can help one become calmer. There are many apps, such
as Calm, which can help one relax by listening to soulful natural soundscapes.

5 Tips to help with Stress

GET SOME FOCUS ON YOUR DO THINGS THAT
SLEEP STRENGTHS MAKE YOU HAPPY PHYSICAL ACTIVITY

Z z.
Getting enough sleep Take time to think about
0
Find activities or Exercise takes our mind It can be hard to manage
helps you grow and what you're good at hobbies that make off stress and releases stress alone. Talk to a
develop normally, stay and ways to do more you happy and chemicals in our brain parent, teacher or other
active throughout the of those things. By incorporate them into that make us feel better. trusted adults about your
day, and maintain overall focusing on and building your daily life. This can be anything problems and they may
health. For teens, this your strengths, you can from a stroll in the park, be able to help you find
means about 8-10 hours keep your stressors in to a downhill bike ride or ways to manage your
ot sleep each night. perspective. a basketball game stress
with friends.

Source: WIHFA.org/teens
There are apps which can help you in managing stress in a better way. One
such app is called Calm, which has natural sounds, guided meditation sessions
and sleep stories.

©Senses
To Help Manage Stress

Look at a favourite photo. Find an area with natural

light. Surround yourself with warm or cool colours. Smell
Enjoy clean, fresh air. Smell your favourite
flowers. Light your favourite candle. Try
aromatherapy to help de-stress.

Indulge in a small piece of dark chocolate. Enjoy a

healthy, crunchy snack. Sip some hot tea/coffee.

Tune in to a nature soundtrack to hear the soothing Wrap up in your favourite blanket. Soak in
sound of running water. Listen to some uplifting a hot bath. Pet a dog or cat. Give yourself a
music. Sing along to your favourite number. neck or hand massage.
jESSBBT
Write and present a talk on “What I wish to become and what I do about it” . Let
your imagination run and discuss with your classmates.
v____________________________________ _____________________________________ )
Friendly Social Network Self-limitations

4*
Reduce the noise

Stress Management

Time Management________ Sleep Well

B3BB30
Preparing a write-up or an essay on experiences during a holiday trip
If you have taken a school trip or fam ily trip recently, prepare an essay on your
experiences during the trip (day-wise). Include what all you did independently and
what you did with friends/fam ily. Share notes with your classmates.
V____________________________ _____________________________________________ /

TYPES OF SELF-AW ARENESS

There are four areas of self-awareness that you should continuously develop:
1. Self-awareness of your strengths
Take time to discover your strengths by paying attention to these things:
• What makes you happy and when do you feel energized?
• What is your natural inclination?
• What activity can you do without getting tired?
2. Self-awareness of your weaknesses
We are not perfect. Everyone has some flaws and weaknesses. It is okay to
acknowledge them and work on them. The important thing is to be aware of
them. The best way to discover your weaknesses is:
• Ask your classmates, teachers, family and friends what skills do they think
you are not good at or what kind of work will they not give you if it is
needed to be done.

ES.22 E ssentials of A rtificial Intelligence - I X

 Self-awareness
Aspirations Self-knowledge

Dreams Identity

Quality of
Life Talents

Human
Capital Potential

3. Self-awareness of your problems

Many a time we develop certain habits which are not necessarily our
weaknesses but we are ashamed to talk about them or we try to hide them.
Think about—
• Any habits which you can’t admit to anyone
• Any acts or activities you do which are not socially acceptable
• Any character flaws which you would want to improve
4. Self-awareness of your emotional triggers
Your anxiety will only make people around you more anxious. So, you need to
discover what situations trigger anxiety in you.
Ask yourself what kinds of situations make you feel anxious. Some of the
situations could be:
• When things don’t happen as per plan.
• When there is no clarity about the future.
• When a situation arises for which you are not prepared.
• When you don’t get the intended result.
• When people lie to you.

S elf -M anagement S kills - I ES.23

Objective Type Questions :

1. Which of these is not an exhibitor of self-management?

(a) Habits (b) Manners (c) Deeds (d) Communication
2. Which of these is not a self-management skill?
(a) Financial strength (b) Self-confidence
(c) Independent working (d) Time management
3. Which of these are physical signs of stress?
(a) Breathlessness (b) Nausea (c) Fatigue (d) Fever
4. Which of these factors has no role to play in building self-confidence?
(a) Social (b) Cultural (c) Physical (d) Financial
5. Which of these is not a benefit of stress management?
(a) Improving mood (b) Promoting longevity
(c) Increasing efficiency andeffectiveness (d) Increasing weight
6. Which of these is not a tip for building self-confidence?
(a) Getting rid of negative thoughts (b) Staying happy with small things
(c) Thinking positively (d) Earning more money
7............................. refers to a "physical, mental or emotional strain or tension".
8............................. Is a conscious effort to prioritize tasks according to time at hand to
increase productivity, efficiency and effectiveness.
9. General neatness, grooming, dressing, apt verbal and non-verbal communication are
components of good personality. (True/False)
10. Prolonged levels of stress are bad for health. (True/False)

Subjective Type Questions •

1. What are the various causes of stress? Explain with examples.

2. List down some tips for building self-confidence.
3. Why is it important to do Strength and Weakness analysis? Can you do it for your
own self?
4. Why do we need to focus on self-management? How does it improve one's life?
5. Can you list down few stress management techniques?
6. Can technology-led innovations help in reducing stress? Explain with examples.
7. What does it mean to work independently?
8. What are some of the physical signs of stress?
9. What is the importance of Personality Management and how can you improve upon it?
10. What are the various types of self-management skills?

ES.24 E ssentials of A rtificial I ntelligence - I X

 orv

a
A 1' ICT S kills-
f

INTRODUCTION
ICT stands for Information and Communication Technology which deals with the
use of electronic devices such as computers, smartphones, etc., to manage data or
information. All the smart devices around us including Al and loT form part of ICT.

Role and Importance of ICT in Our Lives

ICT applications are found in every walk of life—from education to health care,
e-commerce to environment, etc. We use it for shopping, booking tickets, playing
games, watching our favourite movies, listening to music, etc.
Some examples are:
• We can read news online.
• We can get connected with our friends and family on messaging apps and
audio/videoconferencing.
• We can shop online.

ICT at Home
Many of us have personal computers and smartphones at home which can be used to
play games, for listening to music, sending or receiving emails, online banking, etc.

ICT in Education
ICT in education includes development of technology around learning and
development processes. Using online education tools and mobile apps for learning
has become a mainstream methodology in education. It is referred to as Blended
Learning.
The concept of ‘flipped classroom’ is becoming popular where instructional content
is consumed by the students at home in the form of videos and classroom time can
be used for interactive activities, discussions and projects that illustrate the content.

ICT S kills —I ES.25

Some of the advantages of ICT in education are:
• Complicated topics can be easily explained using animations and graphics.
• Use of audio-visual aids makes the class engaging.
• Students can learn at their own pace and at any time, anywhere.
• Personalized learning can be enabled.

ICT in Health Care

Telemedicine is a growing area in health care through which you can take virtual
consultation with doctors anywhere in the world. This practice gained momentum
during the lockdown and has since continued.

ICT is also used in health care research which helps in checking the spread of
infections and undertaking vaccine trials.
Hospitals also use ICT to store patient data and other medical records. Through
simple apps and integration with wearables, doctors and patients can check their
vital information and historical data to assess onset of diseases proactively.
Computer-based machines are used for MRI, CT scan, etc., in diagnostic centres.

ICT in Governance
ICT is increasingly being used to improve efficiency of operations and transparency.
Many government services such as Passport, PAN, Aadhaar, etc., are now available
online. Electronic voting machines (EVM) are used during elections and all the data
is collated in electronic systems. Entire tax filing is online now. You can pay utility
bills online as well while sitting at home.

ES.26 E ssentials of A rtificial I ntelligence - I X

 e-Filing Anywhere Anytim e m A b out Us Feedback A c ce ssibility O p tio n s C ontact Us e-N ivara n

Inum T« Dipirtnwnt, Gownmutl of India

Search

UDIN to r Audit F orm e/A udit certificates can be updated till 30th April, 2021. Last data fo r linking PAN with Aadhaar Is 30th Juns, 202

List of Income Tax Return NO MORE nro-*AOiFO *M New fo e Filing?

Register Yourself
and Forms available for e- P ll t , L I/ ,
Filing FILLING F(
FORMS.
Tax Information and
GET YOUR A r Registered User?
lo g in Here j
Services PAN INSTANTLY ONi AADHAARNUMfifift >

e-Filing Brochures A PP LY N O W It. Forgot Password?

Quick Links News & Updates Download

O Instant PAN through Aadhaar 03/04/2021 N ew O IT Return Preparation Software
User can now download and fill the offline utility for ITR 1 & 4 (AY 2021-22) The
O Link Aadhaar O Other Forms Preparation Software
utility for other ITRs will be enabled shortly
O Tax Calculator FY 2020-21 O DSC Management Software
03/04/2021 N ew
O e-Pay Tax | Challans ITR submission for AY 2021-22 will be made available shortly O e-Filing FAQs on DTVSV Forms
O e-Verify Return O FAQs on Return Filing
03/04/2021 New
O View Form 26AS(Tax Credit) a For AY 2021-22, the ITR1 to 4 can be filled using single JSON Utility Import of O ITR Notified Forms AY 2020-21
Prefill file is mandatory in utility Refer the attachment for more details 0 (1113 KB)
O Outstanding Demand a How to?
03/04/2021 New
O ITR Status
Excel & Java version of ITR utilities will be discontinued from AY 2021-22 O Scan the Supporting Documents
o Know Your TAN | AO
01/04/2021 N ew
O e-Verify Return
O Verify Your PAN Details
New Form 10A is available for e-Filing for registration or provisional registration O Reset Password
O Tax Calendar or intimation or approval or provisional approval
O Update PAN/Aadhaar Details
O Check Refund Dispatch Status
O Report Account Misuse
O Submit Tax Evasion Petition or
M ore Statistics
Benami Property holding
This PorteI Is really useful even though I am n o t an accounting person. Its Individual Registered
O Verification of applicability u/s 194N
really user friendly and needy organised. Hats o f to the Indian Users
Authenticate Qovemmendlncometax Departm ent 01-07-20H Registered and
Aadhaar Linked
O Notice/Order Issued by ITD Not-Registered but
Aadhaar Linked
O Reminders I Campaigns
e-Verified ITRs (20-21)

ICT in Business
ICT is used in businesses too for different purposes such as:
• Enterprise Resource Planning (ERP) for process management and compliance
• Financial Management
• Customer Database Management
• Online payments/Fintech
• Mobile apps and chatbots for customer communication

ICT and Society

ICT has affected society in a big way but it has also led to a digital divide between
people who have access to ICT and those who don’t. This leads to the creation of
tw o categories of population:
1. People who use ICT in their daily lives and for services such as net banking,
e-learning, e-governance, virtual assistants, etc.
2. People who don’t have access to a computing device and/or internet.

ICT S kills —I ES.27

ICT Tools
ICT tools are devices that are used to communicate and share information.
Some of these devices are:

Mobile Phones
Mobile phones are lightweight and portable devices that you
can use to make calls, stay in touch with your friends and
relatives, any time, anywhere. Smartphones enable you to do
web search, send/receive emails and access a range of apps.

Tablets
Tablets are small computing devices with a touch screen. Tablets are
popularly used in many e-learning programs with pre-loaded content.

Radio
Radio is used for information broadcast and entertainment.
Nowadays, there are talk shows as well with RJs on radio.

TV
Tele means distance, vision means seeing. TV means seeing at
a distance. You can watch TV shows, music programs, movies,
etc., on TV.

Online Newspapers
Online newspapers are now available and all the news is at your fingertips.

T h e E conomic T im es
English Edition * | 08 April, 202110:24 AM 1ST | E-Paper

=E H om e Q ETPrime Markets N ew s Industry RISE Politics wealth MF Tech Tote Opinion NRI Panache ET N O W More - Q,

News Live' Markets Uve < Stock Screener MFScreener OM vs New Tax Regime Candlestick Screener Sunday ET Covid Updates ETSOO

V a c c in e s h o r t a g e s b e s e t 13ETPrime
Did Ambuja's veteran independent

fit
I n d ia 's fig h t a g a in s t a
directors ask the right questions?
seco n d v im s w a v e
9 m ins read * By N Sundaresha Subramanlan
Did RBI get it right in the first MPC of FY22?
How a retreating bull market can impact
Few takers for TLTRO despite deadline exlensi an IPO's quality

9 m m s read * B y Sidh artna Shukia

ftir builders, virtual tour is the new normal

How RBI plans to fight the second wave How Tata Group is taking a leaf from an
old book to ace ed-tech
9 m ins read * By Jochelle Mendonca
• Sense* returns to Mt SOK led by bank. IT ttocks
Second w ave of coronavirus How govt can improve tire lending
hits IPL's out-of-home
7 stocks that analysts suggest lor solid returns discipline of state-run banks
ilSP

n
revenues for a six
Why MFfirat called for exiting Templeton funds
3 mins read * By Rama Bhushan

A fte r years o f bloodletting, the

country's biggest telecom rivals
TOP NEWS LATEST NEW S RECOM M ENDED
find peace in unused spectrum

AstraZeneca sends legal notice to SU

i India reports over US0O O new Covid-19 ci

ES.28 E ssentials of A rtificial Intelligence - I X

Email
Email (Electronic Mail) is used to send and receive information. There are a lot
of email clients such as Gmail, Yahoo, etc. which enable you to send and receive
email messages.

BASIC COMPONENTS OF A COMPUTER SYSTEM

A computer is an electronic device that accepts data from input devices, processes
it using processing device and displays output using output devices.

Hardware and Software

A computer system comprises hardware and software.
• Computer hardware: Physical parts such as input devices, output devices,
central processing unit and storage devices are called computer hardware.
• Computer software: Software are the applications that run on computer,
e.g., MS Office, OS, email, etc.

Computer Hardware
The physical components of a computer are known as hardware. A computer
basically consists of the following hardware parts:
• Input unit
• Processing unit
• Main Memory Unit
• Secondary Storage Unit (Auxiliary Storage)
Output unit
CPU
Memory Unit

INPUT Control Unit OUTPUT

UNIT UNIT
Examples:
Keyboard, Mouse, ALU 1. Soft Copy Devices, e.g., Monitor
Light Pen, Joystick, etc. 2. Hard Copy Devices, e.g., Printer, Plotter

Secondary
Storage Device
Examples:
Magnetic tape, Magnetic Disks, CDs, etc. /

t
Peripheral Devices

ICT S kills—I ES.29

Input Devices
Input devices are used for entering data or instructions into the computer,
e.g., keyboard, virtual keyboard, mouse, touch screen, etc.

Central Processing Unit (CPU)

The Central Processing Unit (CPU) is like the brain of a computer. The functions
of the CPU are:
• It controls the sequence of command operations
• It allocates internal command to other systems
• It helps in allocation of memory to processes
It consists of the following main units:
• Arithmetic and Logic Unit (ALU)
• Control Unit (CU)
• Main Memory Unit

Arithmetic and Logic Unit (ALU)

All the calculations and comparisons are done in this unit, which include:
+ (addition)
- (subtraction)
• (multiplication)
/ (Division)
A (Exponent)
ALU also performs the following logical operations:
< (less than)
> (greater than)
<= (less than or equal to)
>= (greater than or equal to)
<> (not equal to)

Control Unit
The control unit controls the data flow from the input devices to memory and
from memory to output devices. No processing is done here.
The interpretation of instructions happens in the control unit and then the signals
are sent to other components for necessary action.

ES.30 E ssentials of A rtificial Intelligence - I X

The next instruction is then selected interpreted and executed. This continues till
all the instructions are processed.

R egisters

These are temporary storage areas found in the CPU section of a computer.

Software
Software is a set of computer programs and applications that performs a particular
task. The tw o types of Software are:
1. System Software
2. Application Software

System S o ftw a re
System software is simply a set of applications that are designed to control and
manage the running of a computer system. Examples of system software include
Operating Systems (OS) and Language Processors.
• Operating System: Operating System (OS) is the main program that runs the
computer. When we boot a computer, OS is loaded on to the computer’s
memory and starts the computer. Some of the popular OS are Windows, Mac,
OS, etc.
• Language Processors: As we know, a computer understands instructions in the
form of binary, i.e., 0 and 1. When we use a programming language, such as
Python, to write code, it is written in high-level language. This gets converted
into machine language in order to execute the commands and for machine to
interpret it. The program that converts instructions in high-level language to
machine-level language is called a language processor. The program which is
translated into machine-level language is called the object program.
Language processor can be of the following types:
■ Assembler: A program that translates an assembly language program into
machine language.
■ Compiler: A program that translates a high-level language program into
machine language, for example, C, C++ compiler.
■ Interpreter: A program that translates a high-level language into machine
language program line by line, fo r example, Python Interpreter.

ICT S kills —I ES.31

 TYPES OF LANGUAGE TRANSLATORS

Computer
Instruction

High Level
Language

Application Software
Application software is a program that is developed to do a particular set of tasks,
for example, a school management software that keeps a database of all students,
teachers and users in the school and maintains their records.
Application software can be classified as general-purpose application software and
specific-purpose application software. Examples of general-purpose application
software include MS Office, etc.
• Specific-Purpose Application Software: Specific-purpose application software
is a program that does a specific task. Examples of specific-purpose application
software are Payroll software, Call Centre software, Lead Management, Financial
Management, etc.

Utility Program
A utility program is used for maintenance work of the computer or its components.
Examples of utility program include file management programs, compression tools,
antivirus software, etc.
• File Management Tools: These tools help in storing, searching and managing
files and folders on the system, for example, Windows Explorer, etc.
• Compression Tools: These help in compression of large folders and files so that
they take less storage space and are easy to store and transfer, fo r example,
WinZip, etc.
• Antivirus Software: This program helps in scanning the system to check for virus
and vulnerabilities. It then helps in getting rid of them as well, for example,
McAfee virus scan.

ES.32 E ssentials of A rtificial Intelligence - I X

• Disk Management Tools: These are a set of tools which are used to manage
disk for optimized use. Some of the tools include:
■ Disk Cleaner scans for files that are occupying space but are not used for
quite some time. It notifies the user that they can delete these files and
create some free space.
■ Disk Defragmenter helps in reorganizing the files and free up space so that
files can be accessed quickly. This optimizes the performance of the disk.
■ Backup is used to make a copy of the data including folders and files on the
computer. In case the original data is lost, the back-up data can be used to
restore the system to an older date with this data.

Primary Memory Storage

Primary memory storage is also called primary memory. It is directly accessed by
the CPU. It can be:
• RAM (Random Access Memory)
• ROM (Read-Only Memory)

Functions of primary memory are:

• Data is fed and held until it is ready to be accessed by other parts.
• Result of the processing is held here along with processing instructions.

Main Memory Units

The elementary unit of memory is a bit. A group of 4 bits is called a nibble and a
group of 8 bits is called a byte.
1 Byte = 8 bits
One kilobyte (KB) = 1024 bytes
One Megabyte (MB) = 1024 KB = 1024 * 1024 bytes
One Gigabyte (GB) = 1024 MB = 1024 x 1024 * 1024 bytes
One Terabyte (TB) = 1024 GB = 1024 x 1024 x 1024 x 1024 bytes
One Petabyte (PB) = 1024 TB = 1024 x 1024 x 1024 x 1024 x 1024 bytes
One ExaByte (EB) = 1024 PB = 1024 x 1024 x 1024 x 1024 x 1024 x 1024 bytes
One ZettaByte (ZB) = 1024 EB = 1024 x 1024 x 1024 x 1024 x 1024 x 1024 x 1024 bytes

ICT S kills —I ES.33

Random Access Memory (RAM)
RAM is used to hold active data and information is accessed in random order; that is
why it is known as Random Access Memory. It is also known as temporary memory
as the information stored in RAM is lost when power supply is switched off.
It is also called read and write memory as information can be read from RAM and
also written on to it. Information stored in RAM may be erased or written over
when the computer is ON.

Read-Only Memory (ROM)

ROM is a part of computer’s main memory. The storage of data and instructions
in ROM is permanent but it can be written over.

ROM is called non-volatile memory as the contents of ROM are not lost when
power supply is switched off, e.g., PROM, EPROM, etc.

Secondary Memory
Secondary memory is used to store large amount of data & information permanently.
The Secondary memory is non-volatile and has larger storage capacity than
Primary memory. Secondary storage devices include CD, DVD, Pen Drive, etc.

^Identify and name the various components of a computer system.

V__________________________________________________________________________ T

ES.34 E ssentials of A rtificial Intelligence - I X

PERIPHERAL DEVICES AND THEIR USES
A peripheral device is an internal or external device that connects to a computer
but does not contribute to the computer’s primary function which is computing.
It helps in supporting a user access the computer.
Different peripheral devices fall into the following three general categories:
• Input devices, such as a mouse and a keyboard
• Output devices, such as a monitor and a printer
• Storage devices, such as a hard drive or flash drive

Input Devices
Keyboard: A keyboard is an input device used to
enter data into the computer. Keyboard can be wired
or wireless. Nowadays, there are USB keyboards.

YOU KNOW?

The first row of alphabet keys contains alphabets Q, W, E, R, T, Y. This is why

it is known as QWERTY keyboard.
A wireless keyboard uses radio frequency (RF) or Bluetooth technology. It has two
parts— a transm itter and a receiver. The transm itter is attached to the keyboard
and the receiver is attached to the CPU.
The different types of keyboard keys and their functions are:
• Alphanumeric keys are used to type alphabets, numbers and special symbols
like $, %, @, etc.
• Special keys, such as Shift, Ctrl, Alt, etc., are used for special functions.
• Function keys, such as F1, F2, F3, etc., are used to give special commands
and these commands may be different for different software.
• Cursor movement keys are used to move the cursor in a document. These
include the arrow keys, PAGE UP, PAGE DOWN, etc.
• Numeric keypad is present on the right of the keyboard and can be used to
enter numbers quickly.

ICT S kills—I ES.35

Mouse: A mouse is used to move the cursor pointer on a monitor.
A computer mouse may be connected to a computer using a USB
port. A mouse usually has tw o buttons with a scroll wheel in the
centre. Wireless mouse is also available now.
There is optical mouse also that uses an LED sensor to detect the
movement of the mouse on the tabletop and sends information to
the computer to mimic the movement on the monitor in real time.

Joystick: A joystick is used to play games on the computer. You may

have seen it in various PlayStation and Xbox gaming devices.

Light Pen: A light pen is a pointing device which is used to draw

directly on the screen.

Graphic Tablet: A graphic tablet is used to create digital drawings

by hand drawing an image on a flat surface, called a tablet, using a
special pen. This special pen is called stylus.

Scanner: A scanner is used to convert a document into a digital

file in real time. You might have used it to scan photographs and
important documents.
Barcode Reader: A barcode consists of a combination of thick and
thin vertical lines found on products. The bar code identifies the
product and the manufacturer. These barcodes can be read by a
device called Barcode reader.
Microphone: A microphone is a device used to record voice and save it
in the form of a digital file.

Output Devices
Computer M onitor: A computer monitor is like a TV screen
and can display anything on the computer in real time. The
size of a computer monitor is measured in diagonal length of
the screen. Computer monitors are available in different sizes.
The monitor has a grid of tiny dots called pixels (We will know
more about pixels in Class X in Computer Vision).
The first computer monitors used Cathode Ray tubes. This was followed by LCD
(Liquid Crystal Display) monitors. Nowadays, LED (Light-Emitting Diode) monitors
are also available.

Touch screen monitors are becoming very popular these days. Have you seen them
in shopping malls and ATMs?
Printer: A printer is a device used to print physical copies of the documents stored
in the computer. The output produced on paper is called the hard copy.
Some commonly-used printers are:
Dot Matrix Printer: It is the most commonly used character
printer which means it can print one character at a time.
Dot matrix printers are also bidirectional, i.e., they print one
line of text from left to right and then the next line from right
to left.

Inkjet Printer: These printers use a continuous stream of ink drops

to print on paper.

Laser Printer: These printers print one page at a time and are
very fast. You may have seen them in various organizations and
shops where print volumes are high.
1 I
Braille Printer: A Braille printer produces patterns of raised dots
of computer files. It works with software that translates the text
to Braille.

Plotters: A plotter is used to print large-sized engineering

and architectural drawings on a paper or a polyester
film.

Speaker: A speaker is used to listen to audio/sound from the

computer. Speakers can be connected with a wire/USB port or
through Bluetooth in the speakers that are wireless.
Storage Devices CDs/DVDs: Compact Disc (CD) and Digital
Versatile Disc (DVD) are optical media for data recording and
storage. To read data from the CD/DVD, we need to have a
CD/DVD Drive in our computer.

ICT S kills —I ES.37

There are tw o variations of CD:
CD-R and CD-RW: In CD-R (Compact Disc-Recordable), data recorded can be read only
and cannot be erased while in CD-RW (Compact Disc-Rewritable), previously recorded
data can be erased and new data can be recorded. DVDs are an evolved version which
can hold much more data than CDs.
A CD can store up to 700 MB of data whereas a DVD can store 4.7 GB of data.

Pen Drive/Flash Drive: A pen drive is a small and

portable device used to store and transfer data. Pen drives
are available in various capacities such as 1 GB, 4 GB,
8 GB, etc.
External Hard Disk: An external hard disk is a
storage device located outside of a computer. We
can connect it using a USB cable. It has a high storage
capacity compared to pen drives. We can use it to
take backups and transfer large files across different
computers. External hard disks have a capacity of around
1 TB and more.

PERFORMING BASIC COMPUTER OPERATIONS

Starting a computer
The steps to start a computer are given below:

1. Press the power button of the computer.

2. When you turn on the computer, lights on the keyboard may blink for a while
and you may hear a sound. The monitor may display a message if a component
is not functioning.

3. The computer starts the operating system in a few minutes. This process of
bringing up the operating system is called booting. The instructions of booting
are built into a chip called BIOS (Basic Input/Output System).
4. When the operating system starts, Windows welcome screen appears. The first
screen that appears after OS is successfully loaded is called desktop.

ES.38 E ssentials of A rtificial Intelligence - I X

5. Now you can use various applications and open files as needed.

Shutting down the computer

Before you shut down the computer, make sure you
have saved all the files and applications. You can
make the computer go to sleep or do a complete
shut down. To turn off the computer completely,
click the Start button and then click Shut Down. If
you choose Sleep option, the system uses very little
power and you can start where you left off.

OPERATING SYSTEM
You have been using a computer and are aware of its vast applications, right? It is
important to know that just like human body has brain and other body parts which
require rest and care, a computer also has CPU/motherboard and other parts which
require constant upkeep and maintenance to run properly. An operating system
is the heart of the machine.
So, what is this operating system?
An operating system is a software that serves as an interface between the user
and computer. An operating system does the following:
• Manages all components of the computer and their status
• Checks whether the components are functioning properly or not
• Controls resource allocation to various applications
• Manages the computer memory (hard disk) and its allocation
• Manages structure of the files and directories
• Manages files (add, edit, delete, move, etc.)

ICT S kills—I ES.39

Types of Operating Systems
There are few commonly-used operating systems:
• DOS (Disk Operating System)
• Windows-An operating system by Microsoft
• L in u x /U b u n tu -A free and o p e n -so u rce so ftw a re
• Mobile operating systems
■ Android (Google)
■ Windows Phone
■ iOS (Apple)
The different types of operating systems are as under:
Interactive (GUI-based): A GUI-based OS is a user-friendly operating system in
which input commands can be easily entered by the click of a button without any
use of machine-level language, e.g., Windows OS.
Single-user, single-task operating system: This type of operating system enables
a user to do a single task on the computer at a time.
Single-user, multi-task operating system: This operating system is used on desktop
and laptop enabling a single user to operate several programs parallelly, e.g.,
Windows OS, Mac OS, etc.
Multi-user: A multi-user operating system lets multiple users to work on the same
computer at different times or simultaneously. This is often used in gaming industry.
Real Time: A real-time operating system is used to control complicated robots and
loT machines. It is a computing environment that instantly reacts to an input by
managing the resources in such a way that a particular operation is executed in
the same period of time every time it is executed, e.g., Windows CE and Lynx OS.
Distributed: A distributed operating system runs on a network of computers. It
integrates various computers in the network into a single mesh of computing
structure, e.g., Windows, UNIX and LINUX.

MENU, ICONS AND TASKBAR ON DESKTOP

Let us see in detail a Windows 7 operating system.

PO PP 44

Windows 7
ES.40 E ssentials of A rtificial Intelligence - IX
Let us revise the use of some of the commonly-used icons.

Taskbar: Taskbar is a long horizontal bar at the bottom of the screen. On the left
of the taskbar is the Start button and on the right appears Date/Time. There are
a few shortcuts as well.

Start Button: It is located on the left of the taskbar. By clicking the Start button, we get:
• Restart: Restart a computer
• Shut Down: Allows the user to turn off the computer
• All Programs: Access to all installed programs and applications. Simply click on
an application to start it.
• Search Box: Search a file or folder
• Control Panel: Change user settings

ICT S kills —I ES.41

Recycle Bln: Files and folders which have been
deleted by the user are stored in the Recycle Bin.
From Recycle Bin, you can retrieve files as well. But if
you do empty the Recycle Bin, they will be permanently
deleted.

1. Start your computer and then shut it down.

2. Name the operating system in your computer.
3. Identify three applications.
4. Name the buttons on the Taskbar.

Manage File Structures

The computer has everything structured as directories, folders and files. File system
is a way to create, store and manipulate a file. There are some rules for naming a
file/folder in terms of characters, length, etc.
Files can be structured into folders and sub-folders for easy navigation.

CREATING AND MANAGING FILES AND FOLDERS

Creating a Folder
The steps to create a new folder are:
1. Double click the Computer icon.
2. Select the drive in which you want to create a new folder, say Local Disk D:.
3. A window will open up showing files and folders in Local Disk D:.
4. Click New Folder on the toolbar.
Or
Right click anywhere in the blank area of the right column. Click on
Select New -> Folder from the shortcut menu.

ES.42 E ssentials of A rtificial I ntelligence - I X

5. A new folder is created with the name New Folder highlighted.
6. Type a name for the folder. Press the Enter key.
View

Sort by
•
Refresh

Pkfflf

Paw* ibcwts-irt
Undo Move CW.2

H r* *

j krw*n m o tutum
m Gadgets i
n
i ff Penonate Is |Ncvvfoldei|

11$ l

Fig. 3.1: Creating a New Folder

Creating a File
To create a new file:
1. Right click anywhere in the blank area on the screen.
2. In the Shortcut menu, click New and select the type of file you want to create.
View » I
Sort by ►I
Refresh

Paste
Paste shortcut

Undo Rename Ctri-Z I

Graphics Properties...
Graphics Options ►I

$ Folder
SO Shortcut

(3) Microsoft Access Database

@3 Contact
[SJ Microsoft Word 97 - MOB Document
v3] Microsoft Word Document
(fjj Journal Document
^*1 OpenOocument Drawing
OpenOocument Presentation
T§| OpenOocument Spreadsheet
■jfj OpenOocument Text
!#] Microsoft PowerPoint 97-2003 Presentation
if] Microsoft PowerPoint Presentation
Microsoft Office Publisher Document
Q Text Document
8) Microsoft Excel 97-2003 Worksheet
fflj Microsoft Excel Worksheet
Jjj Compressed (zipped) Folder
& Briefcase

Fig. 3.2: Creating a New File

Renaming Files or Folders

To rename a file or a folder, the steps are:
1. Right click the file or the folder.
2. From the shortcut menu, select Rename option.
3. Type the new name and press the Enter key.

ICT S kills—I ES.43

Deleting Files or Folders
To delete a file or a folder:
1. Click the file or the folder.
2. Press the Delete key.
Or
Right click and select Delete option from the Shortcut menu.

Restore previous versions

Delete File X
Send to >
Are you sure you want to move this file to the Recycle Bin?
Cut

Copy New Microsoft Word Documentdoc

Type: Microsoft Word 97 - 2003 Document
K Create shortcut
Size: 0 bytes
Delete
LJ Date modified: 08-04-2021 12:32
Rename Availability status: Available on this device
0 ----- Properties

1. Start your computer.

2. Double click on the Computer icon.
3. Select Local Disk D:.
4. Create a folder named "My First Al Document" in Local Disk D:.
5. Rename this folder as "My New Al File".

Copying Files and Folders

When you copy a file or folder, a new file or folder is created along with the copied
file or folder.
You can copy a file or folder as follows:

1. Shortcut Menu:
1. Right click the file or folder you want to copy.
2. Select Copy option from the Shortcut menu.
3. Select the drive or the folder where you want to make a copy of the selected
file or folder.

ES.44 E ssentials of A rtificial Intelligence - I X

 4. Right click and select Paste option from the Shortcut menu.
► Libraries » Documents a pictures ► ▼j * f |j Scorch pktur... p | jo a a '
Organize » £2 Preview - Share with * Print E-mail » -»
a «
» Libranes > Documents ►pretures » »j 4y | [ Search ptetur... f i |
"fa Favorites Docum ents lib ra ry Organize * 3*2Preview v Share with ▼ Print E.mml » e * a ®
Arrange by: Folder *
pictures

Libraries
it Documents library Arrange by. Folder »
Name Date modified Type Size a pictures
0 Documents tot *mu.i/uz OICVIOJXI z.uo rm it c « ullage a,ia» ko

Music S , IMG.1766 8/20/20112:10 PM JPEG image 1,102 KB Name ' D «,m »

13 Documents
£ Pictures S IMG.1769 8/20/20112:14 PM JPEG image 558IC8
9 Videos im 1MG.1771 8/20/20112:15 PM JPEG image 1,163 KB
j f i IMG.1774 8/31/201111:31 PM JPEG image 1416 KB i
19 Computer £»; 1MG.1778 8/31/201111:32 PM JPEG image 1,254 KB
S b IMG.1779 8/20/2011 2:23 PM JPEG image 1,210 KB 1
% Network g ja n j”
Preview ' p ffii
tWGJ 343 KB
Set as desktop background
349 KB
Edit
S.WOJ 177 KB
telM G .l Print _ 1
170 KB

M H j| IMG.1782 Date take Rotate clockwise

M JPEG image Rotate counterclockwise
1

Copy —
Create shortcut
Delete
Rename

Open file location

Properties

2. Drag and Drop Method:

To copy a file or folder, hold down the Ctrl key while dragging it to the new location.

3. Using Keyboard:
1. Select the file or folder.
2. Press Ctrl+C to copy the file or the folder.
3. Open the new destination folder.
4. Press Ctrl+V to paste the copied file.

Moving Files and Folders

When you move a file or folder, the original file or folder is moved to the new
location. This is different from copy file wherein a new copy is created.
You can move a file or folder in the following ways:

1. Shortcut Menu:
1. Right click the file or folder you want
to move.
2. Select Cut option from the Shortcut
menu.
3. Select the drive or the folder where
you want to move the selected file
or folder.
4 . Right click and select Paste option from the Shortcut menu.

ICT S kills —I ES.45

2. D ra g and D rop M e th o d :
To move a file or folder, select the file and drag it to a new location.

3. U sing K eyboard:
1. Select the file or folder.
2. Press Ctrl+X to cut the file or folder.
3. Open the destination folder.
4. Press Ctrl+V to paste the copied file.

Deleting Files Permanently

You can permanently delete the files and folders in the Recycle Bin by the following
methods:
Right click the Recycle Bin icon and then click Empty the Recycle Bin.

Or
1. Double click the Recycle Bin icon.
2. The Recycle Bin window appears.
3. Click Empty the Recycle Bin.

ES.46 E ssentials of A rtificial Intelligence - I X

Restoring Files or Folder from Recycle Bin Window
To recover a file from the Recycle Bin, simply right click the file and click Restore.
Or

Select the file and click Restore this item.

S§ > Recycle Bin O P Search Recycle

Name Original Location Date Deleted

© ch-5.com p_netw orks.doa C:\Users\User2\OneDrive\Desktop\Sheet... 29-01-2021 12:54

I CLASSXHP C:\Users\User2\OneDrive\Desktop\Sheet... 20-02-2021 17:55
# Connector ODBC 5.2.6 C:\Program Frles\Mysql 12-01-2021 17:45
© CS-XII.editedVIVA.doa C:\Users\U5er2\OneDrive\Desktop\Sheet.. 16-01-2021 10:58
© D odder -]p\Usefs\User2\OneOnve\D«sktop 03-02-2021 16:55
□ UveRec. .':\Users\User2\AppData\Roaming\Micro._ 29-01-2021 1254
MySQL I ::\Program Files (x86)\MySQL 12-01-2021 17:49
# mysql-ir I:\Program Files 12-01-2021 17:44
S mysql-sj ":\Program Fiies\My$ql 12-01-2021 17:45
1 mysql-sl :\Program Files 12-01-2021 17:44
© New M icrosoft Excel Woricsheetxlsx C:\Users\User2\OneDrive\Desktop 19-01-2021 17:21
© New M icrosoft Word D ocum entdo-. C:\Users\User2\OneDrive\Desktop 27-01-2021 09fl8

1. Start your computer.

2. Double click the Computer icon.
3. Double click Local Disk D:.
4. Create a new folder named "Projects" in Local Disk D:.
5. Rename this new folder as "My Al Projects".
6. Copy this folder to My Documents.
7. Open folder "My Al Projects" in My Documents.
8. Create another folder named "My New Al Projects".
9. Move the folder "My New Al Projects" to the desktop.
10. Delete the folder "My New Al Projects".
11. Restore the deleted folder.

CLEANING EXTERNAL COMPONENTS

Take the following precautions while cleaning the components:
Computer M onitor: You can use a soft microfibre cloth with a special cleaning
liquid. Do not spray water or cleaning liquid directly on the computer monitor and
always spray it on the cloth first and then gently wipe the screen.
Keyboard: Keyboard attracts maximum amount of dirt between the keys. You can
also clean it by blowing it with compressed air or vacuum. Use a moist microfibre
cloth to clean the top of the keys. You should never spray any liquid cleaner or
water directly on the keys of keyboard.
Optical Mouse: You can clean the bottom of the optical mouse with a microfibre
cloth and cleaning liquid. Lens area can be cleaned by a cotton swab.

ICT S kills—I ES.47

Digital Camera: Never touch the lens of the camera with bare hands or spray any
liquid. To remove dust, you can use soft lens brush and wipe with a soft, dry cloth.

CDs and DVDs: Keep CD/DVD in a proper case to prevent damage. If there is some
dirt on the CD or DVD, it may not work at all. Finger prints and dirt can be removed
by lightly rubbing with a clean lint-free cotton cloth. If the dirt is hard to remove,
you can use special cleaning fluid. Put a drop of fluid on a swab of cotton and
gently rub it.
Besides, keep the computer away from water, drinks and food items. Also, keep
the system covered with a cloth when not in use.

P R O T E C T IN G C O M P U T E R A G A IN S T VIR U S ES
A computer virus is a software program that attaches itself to other programs and
alters their behaviour. A virus may or may not be destructive. Some viruses erase
data or corrupt programs on the computer.

v J

Virus can enter a computer system and multiply within by attaching to various
folders and files. Some viruses may get activated at a certain date and time.
A virus can spread in different ways— through an email or a pen drive. It may also
infect the whole network and spread from one computer to another.
A system can get infected with virus in some of these ways:
• Infected files on the system
• Infected pen drives/CD/DVD-ROMs
• Infected file attachment with emails

ES.48 E ssentials of A rtificial Intelligence - I X

A computer virus cannot do the following:
• It cannot infect files on CD or DVD if they are closed for writing.
• It cannot infect computer hardware like keyboard, mouse, etc.
Some of the common signs of a virus attack are:
• Computer runs very slow
• File names or size changes
• Computer often stops responding or hangs
• Unusual error message appears on the screen
• Computer shuts down and restarts on its own

iS S E S D
1. Discuss the difference between human virus and computer virus.
2. Discuss various signs and symptoms that indicate that computer is infected with virus.
3. Discuss various ways in which computer may get infected with virus.

Scanning and Cleaning Viruses

The following points should be kept in mind to prevent virus infection:
• Install and use antivirus software. You can see some examples of antivirus
software below.

• Keep antivirus software version updated.

• Always scan the files you download from mail or the internet.
• Do not open emails from untrusted sources.
• Don’t allow any untrustworthy person to use your system.
• Don’t use unknown devices such as Pen drives, CD-ROMs on your computer.
• Don’t click on unknown ads or pop-ups while using the internet.

ICT S kills —I ES.49

Some examples of antivirus software are McAfee Virus Scan, Norton Antivirus,
Quick Heal, Microsoft Security Essentials, etc.

For an antivirus program to be truly effective:

• It needs to be running in the background at all times. This may make the
computer a little slow though.
• Keep the antivirus software updated.
• Run full disk scans periodically.
Let us discuss how to use antivirus security to scan the computer system.
1. Click Start -> All Programs -> Antivirus Name (name depends on what you are
using). Here, we are using Quick Heal Antivirus.
Or
Double click the Antivirus icon on the desktop.
2. The main window of the Antivirus system opens.

p | Quick Heal Total Security D = -

P ro d u c t G u id e
x
Virus database 17 M arch 2021 License: 183 day(s) left

System is secure

Total Trackers Blocked

o o o <3*
All Time Last 30 days Today

News
Can connected systems lead to the take-over of
y.<?u.r car?find wt{
PCTuner Scan Options
Show Al l

© 1994-2020 Quick Heai Technologies Ltd. & %> [1

To perform full system scan:

1. Click Scan Options -» Full System Scan.
2. The scan process will begin.
3. On completion of the scan, the scan statistics and reports are provided.
4. Click Close.

ES.50 E ssentials of A rtificial Intelligence - I X

 E l Quick Heal Total Security Product Guide ) — — X

< - Performing full system scan

Scanning in progress...

F o ld e r : C:\P rogram Files\W indowsApps\king.com .CandyCrushSaga_1.1970.1.0_x86_kgqv...\anim ation s

E x t r a c tin g :

S ta tu s j A c tio n

Files scanned 26756 Files repaired 0

Archive/Packed 252 Files quarantined 0
T hreats detected 0 Files deleted 0
DNAScan warnings 0
B o ot/P artition viruses 0

Scanning status 5%

0 Shyt down PC when finished

Skip Folder Skip File Pause Stop

Other antiviruses like McAfee have a similar method to scan full computer system
for viruses and remove them.

1. Name the antivirus software installed on your computer.

2. Scan your computer with the antivirus software installed on your computer and
share the scan report.
\__________________________________________________________________________ 7

ICT S kills —I ES.51

Removing Temporary Files
Temporary files are created when you are running computer programs.

Windows programs create a .TMP file as a temporary file. Temporary files are also
created by browsers to store your browsing history.

To Remove Temporary Files:

1. Double click Computer icon on the desktop.

2. The Computer window opens.
3. Right click Local Disk C: and select Properties option from the Shortcut menu.
4. The Properties window opens.
5. Click Disk Cleanup.
6. The Disk Cleanup for C: window appears.
7. Click the Check boxes next to Temporary Files, Temporary Internet files, etc.,
that you want to delete.

8. Click OK.
9. A confirmation message will appear.
10. Click Delete Files.
11. Windows will delete all the temporary files on the computer.

ES.52 E ssentials of A rtificial Intelligence - I X

Removing Files of Temporary Folder
The steps to remove files of temporary folder are:
1. Press Windows button+R on the keyboard.
2. The Run dialog box appears.
3. Type %temp%.
i j j j f lm i
* _ ____

g * Type the name of a program, folder, document, or Internet

resource, and Windows will open it for you.

Qpen:

'*5? This task will be created with administrative privileges.

[~ OK ] \ Cancel Browse...

Fig. 3.3 : Run Dialog box

4. The Temp folder opens.

5. Click Ctrl+A to select all the files in the folder.
p u y u u p 'j

^ | i N e tw o rk it p lu g tm p - 4 12/12/2011 3;50 PM File folder

M TCDlA46.tmp 10/14/2011 3:51 PM File folder
it TC01A57.tmp 10/14/2011 3:51 PM File folder
M TCDlA97.tmp 10/14/20113:51 PM File folder
U TCDlAB7.tmp 10/14/20113:51 PM File folder
J t TCDlAB8.tmp 10/14/20113:51 PM File folder

Fig. 3.4: Selecting files in Temp folder

6. Press Delete key.

7. A message box appears. Click Yes to confirm.

ICT S kills —I ES.53

 8. All the files in the Temp folder will be deleted. Some files which are related to
current open files may not get deleted.

^Can you remove all the temporary files from your computer? Capture screenshots
and share with your classmates. Also note what the size of all the files was and how
much did it take to delete them. j

INTRODUCTION TO INTERNET
Internet consists of an integrated network of networks. It is a medium of
communication and exchange of information between different devices. When
we use internet for general browsing/searching, it is called web surfing.

Important Terms Related to Internet

w w w (World Wide Web): www stands for World Wide Web— a network of
computers.
Protocol: Protocol is a set of rules followed by the computer while communicating
or transferring data on the internet to other computers.
HTTP (Hyper Text Transfer Protocol): This protocol defines the rules to be
followed while transferring information. The information may be in the form of
text, images, videos, etc.
You might have seen this in most URLs.

https is the secured version of http. It has a security certificate as shown below:
Safari is using an ancryptad connection to www.googla.com.
Encryption with a digital certificate keeps information private as it's sent to or from the
https website www.google.com.

G3 GlobalSign
U S G TS CA 101

www.googla.com
Issued by: G TS CA 101

> T ru s t

> Details

? Hide Certificate

ES.54 E ssentials of A rtificial Intelligence - I X

URL (Uniform Resource Locator): Each web page has a unique address called the
URL that identifies its specific location on the network.
The URL has tw o parts:
• Protocol Identifier: It identifies the name of the protocol used.
• Resource Name: It specifies the complete address to the resource on the
internet.
For example, http://www.mywebsite.com
Here,
http is the protocol
www.mywebsite.com is the resource name

Applications of Internet: Some of the applications of internet include:

• Email
• Audio/Video Conferencing
• Social networking
• E-learning
• E-commerce
• Fintech (Financial Technology)
Email: Email means electronic mail. It is a message in electronic form sent using
internet.

COPPIR . x
= M Gmail :H ^ E m ail C o n ta c ts

RECIPEINTS IN COPPER

<- • • • £ Kaitttn Smith

AcccmtmExocutte*at Mtwco

New Project 1 Scan Jon es

9
P»txtw<t Muv* «r at H m c o

Sarah Smith

t me, Kaithn. Sean. Natalie, Sam, Morgan, Prtanka

t i l — IIHPTM...1—
©
RECIPIENTS NOT IN COPPER

N atalie H offm an

CMIHMHM— Sam M cLean

smc»**n#fl»«rco.com

— I o M organ Lee
mloownewco.com
tH —
HUM
A Prianko Suth
ptute<v n*v#co com

ICT S kills—I ES.55

Audio/Video Conferencing
It is an audio/visual communication between tw o or more persons at different
locations.

Social Networking
Social networking helps to connect with our friends and family online.

ES.56 E ssentials of A rtificial Intelligence - I X

E-learning

E-shopping
You can do shopping online 24x7 with e-commerce websites and pay digitally.
d I? ■ f <S> N ew Delhi 110063 |

EE All Fresh Best Sellers Electronics Today's Deals Buy Again Amazon Pay KartIk's Amazon.in Mobiles Prime * Customer Service [ Shopping m ade easy | Download th e app

Dj
OnePlusGR50
S ta rtin g ■ '39,999
Prime Earty Access I Sale starts
14* April. 12PM ! 15"-April, 12PM

Hi, K artik Trending now cooling Top p icks fo r your home Sho p on the Am azon App
Customer since 2015
appliances | Up to 4 0 % off 1* Fast, convenient and secure |Over 17 crore
products In your pocket
Top links for you

9
Download the Amazon App

m
LAPTOPS
FROM TOP BRANDS

s»msm
Single doors |Starting Double doors |Starting

ICT S kills —I ES.57

Internet Browser
An internet browser is a program that is used to load websites for end-user
(client side). It is the common interface between server-side and client-side. Some
commonly-used web browsers are— Google Chrome, Internet Explorer, Apple
Safari, etc.

W Safari file Edit View History Bookmarks Develop Window Help o m v ©

• • • (D < > 1 a google <:om
1 © © + IQ

1f
<
Q.

Google Search rm Feeling Lucky

Google offered in: ffc# S od ) ft *<£)$ JjWkfl aexajfi* Vsnft

India

About Advertising Business Mow Search works Privacy Terms Settings

_______________________________ j

Fig. 3.5 : Safari Browser

Websites and Web Pages

Website: A website is a collection of individual web pages connected through
hyperlinks and sitemap.
Web Page: A web page is an individual page of a website. It is written in Hypertext
Markup Language (HTML)/Cascading Style Sheets (CSS).

There are tw o types of web pages:

• Static web pages wherein all the information is presented to the user exactly
as it is stored.
• Dynamic web pages wherein a web page is controlled by an application and
displays different content each time it is viewed. Different users will view the
page differently based on their navigation and access privileges. For example,
when you create your account on social media, you will see different content
on the page depending on your profile and activity.

Home Page: The first web page of a website is called the home page which is the
main page. Generally, clicking on home button on the website will take you to the
home page.
BENCHMARKS CU»£0 |(S#V—f NSfcGAINER-LARCECAP FEATU RED FU N D S AAAA A S Y R ETU RN
N ift y 14,683.50 T 4 5 .7 0 A d a n i P o r t * S E Z 835 55 T 93.30 ICICI Prudential L a r g e Tu js % [tHvs«Tnowl market watch v

K a rtik •

T h e E conomic T im e s QETPrim*

Home button E n glish E d itio n » | 0 6 A p ril, 2021.04:32 PM 1ST I E-Paper

\
Home 13 E T P rim e M a rk e ts News I n d u s try RISE P o litic s W e a lth ME T ech Jobs O p in io n NR I Panache ET N O W M o re • Q
N ew s Live! M arke ts Live I S tock Screener M F S creene r O ld Vs N ew Tax R e g im e C a ndle stick Screener S u nday ET Covld U p dates ETSOO

How to avoid common 13ETPrime

B yte D a n c e a c c o u n t fre e z e b y In d ia b a c k in
but costly mistakes n e w s , fre s h b a t t le b r e w s
when buying insurance 3 m in s read • B y Sachin Dave

T h e m o m e n t w h e n In d ia n H N I s m in d s e t c h a n g e d
C o v id is h a u n t in g m a rk e ts o n c e ag a in
W h a t s h o u ld M F In v esto rs do?
N ig h t c u r f e w im p o s e d in D e lh i t i ll A p r il 3 0

A m in s read • By Shivani B azar

N o c h a n g e in p e t r o l d ie s e l p ric e s fo r fu ll w e e k

O il c a n 't b e o n A t m a n ir b h a r list: b la m e
C a n y o u s u e i f y o u r M F d e v ia t e s fr o m s ta t e d plan ?
a g e in g w e lls , l a x d is c o v e ry

8 m in s read • B y P rashant M u k h e rje e

M e t a l p h a r m a s tk s h e lp S e n s e s e k e o u t g a in
From M eerut's narrow lanes H o w te ch n o log y, fr a u d -d e te c tio n m o d e ls

K E R ra is e s re c o rd SIS b illio n fo r A s ia fu n d emerges an unlikely winner a r t m a k in g N F C p a y m e n t s s a fe

o f US-China trade w ar 6 m in s read • B y A n a n d 3

P V R in o x L e is u r e f a l l a m id n e w C ovid -1 9 c u rb s
15m ins read • By Neha Dewan
The man who changed stock
trading in India could soon change C O R O N A V m U S C A M S | Live upd a te s
TOP NEWS LATEST NEWS RECOMMENDED
how India invests in MFs, too
I n d ia * W o rld
A ir t e l s e lls 8 0 0 M H z s p e c tr u m to Jio
• C la s h e s i n W B ; v o tin g p e a c e fu l in T N , Ker, A s s a m

ICT S kills —I ES.59

Email Applications
Email has the following benefits:
• Easy to use and access any time, anywhere.
• It is fast to send a message— emails get exchanged within seconds.
• It can be used to transmit text, images, audio, video and other file formats
as well.
• Most email accounts are free to use.
• You don’t need a computer/laptop and can access it from phone as well.
• Eco-friendly option as no paper is used.
• You can communicate with multiple people at the same time.

Email Accounts
You need to first have an email address to send an email.
Email Address: An email address has tw o main parts:
• User name
• Host name
These tw o parts are separated by @ symbol.
For example, see the email ID: [email protected]
In the given email address, kartik.sharma is the user name and dcodeai.com is the
host name.

Email Program: An email program enables you to send and receive email
messages.
Some of the popular email programs are:
• Gmail.com
• Yahoo.com

ES.60 E ssentials of A rtificial Intelligence - I X

Sending and Receiving Email
• To send an email, open your email account with login credentials.

Google

Welcome
£ ) shruti.sharma@ gmail.com v

Enter your p a s s w o rd -------------------------------------------------------------------

.............. -I
0 Show password

Forgot password? Next

• Click Compose option.

= Gmail

|" Compose

□ In b o x 47,049

★ Starred
Q Snoozed
M Important
> Sent
B D ra fts 97

f lt Categories
S o c ia l 1,809

O U p d a te s 47.379

C| F o ru m s 23

• There are few options while composing an email.

New M essage

To | Cc Bcc

Subject

Sans Serif * tT * B 7 U A ■

A » c-3 © a I fe ;

ICT S kills—I ES.61

INTRODUCTION TO SOCIAL MEDIA
Social media refers to various online channels for communication such as Facebook,
Linkedln, Twitter, etc.

A social networking platform is an online community wherein people can

communicate with each other and share messages, photos and videos. They can
collaborate on various social initiatives as well.

Benefits of Social Networking Websites

The benefits of social networking websites are:
• You can connect with anyone around the world at the click of a button.
• Connecting with old friends and school/college alumni becomes easy.
• Catch up with latest trends and updates.
• Advertise and publicize your products and services.

Disadvantages of Social Networking Websites

• These sites expose people to a lot of information which may or may not be
authentic.
• If you share your current location over social media, you can easily become a
target.
• Peer pressure and cyber bullying are also important issues.
• Online interactions have now substituted face-to-face interactions and this has
reduced social skills.
• Social networking has become a distraction for most people. Students who use
social networking too often have lower grades.
• Using social networking for long hours in a day can lead to a sedentary lifestyle.
• Spread of fake news is another major disadvantage.

ICT S kills —I ES.63

How to Stay Safe on Social Networking Websites
You can follow the tips given below:
• Don't post your personal information including photo, contact details, etc., on
social media.
• Manage your privacy settings.
• Don't share your password with anyone.
• Don't send and accept friend requests from unknown people.
• Always log out if you log into a new device.
• Don't chat and send any personal images, etc., to unknown people.
• Don't engage in any arguments or abuse/bullying online.

Blog
A blog is a website which is maintained by an individual. The person who creates
and maintains a blog is called a blogger.
|!} Sorting, se*rc> X j 0 Blogger.com X + — O X

<r -> C ft blogger.com/about/?bpli=1 ft /fa % @ •

8> Blogger SIGN IN

Publish your passions, your way

C re a te a u n iq u e a n d b e a u tifu l b lo g . It's e a s y a n d fre e .

CREATE YOUR BLOG

P Type here to search O HI ^ g* £> « * fa 0> ENG J |)

Twitter
Twitter is a social networking service that allows you to send short messages
to communicate with your friends or followers. You can send short messages of
up to 140 characters, which are called tweets. Twitter is used by people to inform
others about the latest happenings in their lives.

ES.64 E ssentials of A rtificial I ntelligence - I X

 H3 Sortmg. sea X | i f Twin«r. It's X + a x
-> G § twltter.com/?laiig~erHn
a <t 4*0

Happening now
Join T w itte r to d a y .
m

About Help Center Terms of Service Privacy Policy Cookie Policy Ads info Blog Status Careers Brand Resources Advertising Marketing Twitter for Business Developers Directory Settings © 2021 Twitter, Inc ]
0 P Type here to search O Hi e § e> m 1*44
08-04-2021 %

Facebook
Facebook is the most popular social networking site. To use Facebook, you need
to first create an account and then add other users as friends. You can exchange
messages, pictures, etc., with your friends on Facebook.
Companies also set up their own Facebook pages to connect and communicate
with customers and clients.
EDSorting, X j fj Sign ut x □ X

<r -> G § facebook.eom/r.php ☆ 4* 0 i

Email or phone Password
a
facebook ■ mmmm

Create a new account

It's quick and easy,

j Fjrstname fej Surname

; Mobile number or email address

New password

Date o f birth

8 v j Apr v ] 2021 v ®

Gender

O Female O Male O Custom ©

By clicking Sign Up. you agree to our Terms. Data Policy and
Cookie Policy. You may receive SMS noWicalions from us and
can opl out at any time

S ig n U p

8S P Type here to search O B! f d>)) ENG

13146
08-04-2021 *

ICT S kills —I ES.65

YouTube
YouTube allows people to watch and share self-created videos. Companies are also
using YouTube to launch and advertise their product.

BESTOFYOUTUBE

© —
O Spon*
Q flaming
Q MOVlM

Q Nwi

0 FaihlOO S, Beauty
Q Learning

Q Spotlight

Q 360’ Video

O Stows* channels

MOREFROMYOUTUBE

□ P Type here to search

WhatsApp
WhatsApp is a free messenger app for smartphones. WhatsApp uses internet to
send text messages, pictures, audio or videos. You can make an audio and video
call to any person having WhatsApp in their smartphone in any part of the world.
You don’t have to pay anything extra for sending messages or making calls. You
just pay for the internet connection.

ES.66 Essentials of A rtificial Intelligence- I X

Digital India
Digital India is a campaign launched by the Government of India in 2015 to empower
e-governance digitally. It also includes schemes to promote development of
technology in India.
SO Sortii X J f Digit X +

-> C A digitalindia.gov.in

Contrast: High | Normal

1 $

CYBERCRIME
Cybercrime refers to any crime done using computing devices with the internet.
Some examples of cybercrime include:
• Cyberbullying: It refers to harassing someone online by using offensive
and/or threatening language
• Cyber Stalking: It refers to mentally harassing an individual/organization. This
may include making false accusations, defaming, etc.
• Hacking: It means the act of illegally trying to access someone's software
and/or data. The person who is involved in hacking is called a hacker.
• Phishing: Phishing refers to someone trying to steal personal information or
access credentials.
• Spamming: Spam means large volumes of unsolicited emails sent to a large
number of users.

ICT S kills—I ES.67

Cyberbullying: Cyberbullying refers to an event wherein a person uses any
computing device to post content to hurt or harass an individual or organization.
Cyberbullying can occur through SMS, Text and messaging apps, or on social media
where people can view, interact or share content.
It can lead to mental health problems such as anxiety, depression, etc.

Cyberbullying in
social media
Share of social media platforms where
cyberbullying occurs the most: Instagram,
Facebook, Snapchat, WhatsApp, YouTube, and
Twitter

1 Instagram

2 Facebook

3 Snapchat

4 WhatsApp

5 YouTube

6 Twitter

9 "rsiste

Cyber Addiction: We all use internet for social media and e-commerce on a daily
basis, but many a time people get into the habit of overusing it.
Addiction is a mental health problem involving compulsive behaviour. Such a
problem associated with excessive use of internet is known as cyber addiction.
Factors Leading to Cyber Addiction
• Spending less time with family and friends
• Living alone
• Addiction to social media
• Watching movies and online TV shows

ES.68 Essentials of A rtificial Intelligence- I X

Symptoms of Cyber Addiction
• No track of time spent on the internet
• Depression and anxiety while being online
• Getting irritated if someone tries to intrude into your computer time
• Feeling of guilt over use of internet

Effects of Cyber Addiction

• Headaches/Backache
• Insomnia
• Distancing from family and friends
• Dry eyes and eye problems
• Feeling of depression, irritation

: Objective Type Questions :

1. Which of these is not a browser?

(a) Chrome (b) Android
(c) Edge (d) Safari
2. Which of these is not a mouse action?
(a) Single click (b) Double click
(c) Right click (d) Enter
3. Which of these are symptoms of a computer affected with virus?
(a) Computer runs very slow (b) Computer restarts on its own
(c) Computer stops responding (d) Computer can't open a deleted file
4. Which is the place where all deleted items are stored in a computer?
(a) Recycle Bin (b) D Drive
(c) C Drive (d) Desktop
5. Which of these is not an Operating System?
(a) DOS (b) Linux
(c) Android (d) SAP
6. Which of these are two parts of an email address?
(a) Host name (b) User name
(c) SSL (d) http

ICT S kills—I ES.69

 7. Which of these is not a benefit of using emails?
(a) Fast and easy to use
(b) Cheaper way to communicate
(c) Not eco-friendly
(d) Can support attachments in the form of various file formats?
8 .......................... are small files stored on a user's computer and designed to hold data
specific to a particular client and website.
9 .......................... refers to different online communication channels that are dedicated
to community-based input and interactions.
10. Protocol is a set of rules followed by the computer while communicating or transferring
data on the internet. (True/False)

: Subjective Type Questions •

1. What are the different types of operating systems? Which one do you feel is the best
for personal use at home?
2. List down the steps of starting a computer.
3. What is the importance of ICT in personal life and at workplace?
4. Identify various input, output and storage devices.
5. What is a computer virus and why is it harmful?
6. What are peripheral devices? Explain with examples.
7. Explain the benefits of internet. Also, name the various internet browsers.
8. What are the external and internal components of a computer system?
9. What are temporary files and how are they created?
10. What is the use of social media for education?

ES.70 Essentials of A rtificial Intelligence- IX

 o

1 ( I; A
Entrepreneurial
W 3 S kills-I

INTRODUCTION
Entrepreneurship has been there since ages but has been glamourized with the new
start-up ecosystem. In this chapter, we will learn the basics of entrepreneurship,
skills and characteristics of an entrepreneur as well as career options before an
entrepreneur. In fact, many myths and fears surround entrepreneurship which act as
a roadblock for many people to start up. We will also look at how entrepreneurship
positively impacts the society and has immense potential to bring about a change.
Entrepreneurship is the process of setting up one's own business as distinct
from pursuing any other economic activity, be it employment or practising some
profession. The person who sets up a business is called an entrepreneur. The output
of the process, that is the business unit, is called an enterprise.
It is interesting to note that the term “entrepreneurship" has French origins and
is derived from the verb “entreprende”, meaning “to u n d e rta k e ” , and was never
related to economics but to undertaking of military expeditions.
The term “entrepreneur" was first introduced in economics by the
early 18th century French economist Richard Cantillon. He formally
defined the entrepreneur as an “agent who buys means of production
at certain prices in order to sell the produce at uncertain prices in
the future." Richard Cantillon
t ............................................................. ........................................................................... ......................... ...................................................................... ................................................................................. .. ................................................... | H

In economics, entrepreneurship is defined as a systematic, purposeful and

creative activity of identifying a need, mobilizing resources and organizing
production with a view to delivering value to the customers, returns to the
investors and profits for self in accordance with the risks and uncertainties
associated with a business.

Entrepreneurial S kills- I ES.71

 TYPES OF BUSINESSES
# A business is an organization that uses economic resources or inputs to provide
goods or services to customers in exchange for money or other goods and services.
Business organizations come in different types and forms. There are mainly three
types of businesses:
1. Service Business: A service type of business provides no products as such.
They offer professional services, skills, expertise and advise. Some examples
of service businesses are law firms, hair salons, banks, accounting firms, etc.
^ * 2 . Manufacturing Business: A manufacturing business buys raw materials and
transforms it with a process into a new output It combines raw materials,
labour and factory overhead in its production process and creates goods.
3. Hybrid Business: Hybrid businesses are companies that may be classified in
more than one business type. A health care speciality hospital, for example,
combines few raw materials and may create a product (manufacturing),
sells medicines and health care products (merchandising) and does operations/
nursing (service).

M a n f liiw
■Prepare posters of business activities found in cities/villages, using pictures.

ES.72 Essentials of A rtificial Intelligence- IX

 roa
^Discuss the various types of activities generally adopted by small businesses in a
local community.

v _ ................... ................................................................................................................................ ....... ....... ....................................................................................................................................................................... V

M TSMCT61
^Prepare a list of businesses that provide goods and services in exchange for money.

v _ .............. J

ENTREPRENEURSHIP AND SOCIETY

Entrepreneurs are all around us. When you go for shopping, the shopkeeper is also
an entrepreneur who may not have made any sophisticated business plans or raised
venture capital but has set up a business with his savings or debt and is running it.
The relationship between entrepreneurship and society is circular. Entrepreneurs
produce goods and services for society while society, in turn, provides resources to
the business and as a community utilizes those goods and services. Entrepreneurship
benefits society in many ways such as:
• Improving economic growth engine
• Promoting creativity and innovation
• Creating employment opportunities
• Contributing to health and wellness of society
• Helping solve problems of the society
• Contributing to safety and security of the people

Entrepreneurial S kills- I ES.73

Society, in turn, helps boost entrepreneurship by:
• New demand generation
• Producing raw materials
• Ecosystem and community support
• Capital supply
• Infrastructure development
Table 4.1: Difference between Entrepreneur and Entrepreneurship

Entrepreneur Entrepreneurship
Person Process
Organizer Organization
Innovator Innovation
Risk-bearer Risk-bearing
Motivator Motivation
Creator Creation
Visualizer Vision
Leader Leadership
Imitator Imitation
(Source: Khanka, 2 0 12)

Table 4.2: Entrepreneurship and Economic Development

Agrarian Early Mature Service

industrialization industrialization
Level of Entrepreneurial Activity Low Medium Medium High
Small Business Formation Low High Medium Low
Entrepreneurship and Innovation Nil Low Medium High
(Source: M S -9 3 : M anagem ent o f S m a ll a n d M edium E n terp rises, IG N O U , 2 0 1 7 )

Some common entrepreneurship activities related to society include:

Social/lmpact Entrepreneurship: We will talk about use of Al in solving UN SDGs
in subsequent chapters. Social entrepreneurship relates to creating sustainable
solutions for social problems, such as climate change, poverty, health care,
education, etc. The focus of these ventures is not only profitability but larger
social impact.
Agricultural Entrepreneurship: Agricultural Entrepreneurship deals with production
and distribution of agriculture-related products. There have been a number of
innovations in irrigation, farming equipment and pesticides management.

ES.74 Essentials of A rtificial Intelligence- I X

Women Entrepreneurship: Women entrepreneurship is about women organizing
all business activities. The Government of India defines women entrepreneurship
as “an enterprise owned and controlled by a woman having a minimum financial
interest of 51 per cent of the capital and giving at least 51 per cent of the employment
generated in the enterprise to women.”
Micro, Small & Medium Enterprises (MSMEs): Micro, Small & Medium Enterprises
are primarily engaged in the production, manufacture, processing or preservation
of goods and commodities. The government has been laying a lot of emphasis on
MSMEs in recent years.

NEW LOOK MSMEs

Investment in Plant & Machinery Turnover
Micro not more than T 1 crore not more than ^ 5 crore
Small not more than ^ 10 crore not more than ^ 50 crore
Medium not more than ^ 50 crore not more than ? 250 crore
(Source: M in istry o f M S M E )

Innovative Imitative

Innovation + Innovation -
< >

Fabian Drone
R isk -
>r

ROLE OF ENTREPRENEURSHIP
Capital Formation: Entrepreneurship creates value and wealth in society as
entrepreneurs promote capital formation by channelizing financial resources of
the public.
Employment Generation: Entrepreneurship can generate large-scale employment
in society. The employment opportunities are at all levels— company employees,
vendors, partners, etc.

Entrepreneurial S kills- I ES.75

Balanced Regional Development: Entrepreneurship can also help in setting up
economic activity in regions that are backward from a growth perspective. The
government also provides incentives and subsidies for such schemes to promote
overall balanced growth across the country.
Wealth Distribution: Entrepreneurship helps with equitable redistribution of wealth
in society across various geographical areas leading to more equitable and inclusive
growth.
Industrialization and Development: Entrepreneurs act as generators of economic
activity in the society. One enterprise leads to setting up more units as vendors
are needed to support it and slowly the entire ecosystem is set in motion.

REWARDS OF ENTREPRENEURSHIP

Building Large Enterprises: Entrepreneurship results in creation of organizations

with physical, financial and human resources. It is rewarding to create an organization
which can grow into a legacy. For example, Tata Group has been in business for
over 100 years.
Leadership Development: Entrepreneurship helps you develop your leadership
abilities and decision-making skills. You need to constantly take decisions weighing
various options and develop your team. People management is a key skill of
an entrepreneur.
Improved Standards of Living: By generating more employment opportunities and
money flow, general standard of living of the people improves. Goods and services
produced by an enterprise are consumed by members of the public.
Economic Development: Entrepreneurship helps in wealth creation and fresh
capital inflow in the ecosystem. This not only leads to sectoral growth in which
the business operates but to overall economic development of the country.

ES.76 Essentials of A rtificial Intelligence- I X

 Prepare case studies by successful Indian entrepreneurs on their experiences and
success stories.

Qualities of an Entrepreneur
Hard-working: Starting a business requires a lot of hard work and entrepreneurs
always push themselves to achieve more. They have a lot of energy to get things
done before time.
Optimistic: Entrepreneurship can be tough and overwhelming with lots of ups and
downs. Entrepreneurs need to remain positive and always motivated to achieve
success even when things are not going their way.
Independent: Successful entrepreneurs work independently and lead teams with
their vision and passion. They collaborate with teams but lead from the front.
Enthusiastic: Entrepreneurs are always high on energy and enthusiastic about
future prospects. They also like to spread positive energy around and motivate
their teams.
Confidence: An entrepreneur decides to build a product or enter a market in which
most of the times he/she is the first entrant. It requires a lot of self-belief and
confidence to be able to do that.
Perseverance: Many a time, the original idea of the start-up doesn’t work. So, the
entrepreneur needs to pivot and change his ideas to make it work. It requires a
lot of patience and a never-quit attitude.

Functions of an Entrepreneur
An entrepreneur needs to get a lot done in a day— from planning to fund-raising
to execution. Let us take a look at the various functions of an entrepreneur.

Entrepreneurial S kills- I ES .77

 Functions of an Entrepreneur
'■» Responsibility towards the society mm Acceptance
«• Decisiveness - Culture of contribution
'«■ Develop inventiveness - Clarity of target
Experiments - Finding niches
- Be organized - Building networks
- Spread the learnings - Taking initiative
- Control of situations MB Organizing resources
Riskiness for venture . .. H » Identifying opportunities and prospects
To manage the roles MB Innovation
Transfer of learning - Fostering autonomy
Transfer of technology and its acclimatization - Social responsibility
Economical development - Public relations
Division of incomings or earnings - Experience sharing

The Three Archetype M odel: Manager - Entrepreneur - Leader

Manager: Entrepreneur: Leader:
Focusing on Focusing on Focusing on
Complexity Opportunities Change
Plan, Organize, Identify, Innovate, Set direction,
Coordinate, Optimize, Value creation Align and motivate
Control people

Role and Importance of an Entrepreneur

Coordinator: An entrepreneur has to coordinate for the best possible output given
the set of inputs/raw materials at an affordable cost.
Agent: Entrepreneurs are also perceived as "Agents of Change". They are quick
to identify opportunities and convert them into real business and make profits.
Innovator: Entrepreneurs take risks and innovate. They launch a new product or
enter a new market or do both. It may be a tough category or a difficult problem
to solve but an entrepreneur believes in his/her ability to address it and puts in
the required hard work to get it done.
Risk Management: An entrepreneur needs to manage the risk-reward equation
and ensure that all risks are adequately managed without too much strain on the
company’s cash flows. All new rewarding opportunities carry inherent risks and
successful entrepreneurs know how to take calculated risks.

ES.78 E ssentials of A rtificial Intelligence - I X

Taking Inspiration from others: Successful entrepreneurs take inspiration from
other industries as well as economies. A lot of entrepreneurs take ideas from
Silicon Valley and create a local version of the product and make it contextual.
Capital Formation: Entrepreneurs need to raise capital and create wealth in the
process of running businesses. One of the key objectives of a corporation is to
generate shareholder returns on capital. The business needs to be a net contributor
to the economy.
Balancing Role: Entrepreneurs also help in the development of a region through
capital redistribution or opening up MSMEs in various parts of the region and
generate local employment. The government also offers incentives to companies
to invest in certain regions for redevelopment of the area.
Status Transformation Role: Entrepreneurs help in improving the overall standard
of society in terms of wealth, respect and social status. The transformation happens
over the years but an entrepreneur needs to consistently focus on redistributing
wealth among all the stakeholders in society.
Employment Generation Role: Establishment of business corporations leads to
generation of employment opportunities at various levels. It also helps in developing
jobs at the vendors’ as well as the partners’ end.

Why Entrepreneurship
Entrepreneurship as a career option: A lot of universities are now promoting
incubation centres and start-up accelerators to provide entrepreneurship as a
career option. Many students, fresh out of university, are opting for a start-up
rather than going for a regular job.
There is also the new gig and freelance economy wherein professionals don’t run
large enterprises but they work for themselves and are solopreneurs.
It is important to note that while being an entrepreneur can be a lucrative career
option, it is not for everyone and you need to make sure that you have the right
skills and financial support system to pursue it.
Develops confidence and independence: Entrepreneurship can help in developing
the skills of confidence, thinking out of the box and self-reliance. Entrepreneurs are
job creators and stimulate the economy. Some people say that entrepreneurship
is also the best education experience as one gets to learn. This includes learning
about core business areas such as finance, sales, marketing, management and

Entrepreneurial S kills- I ES.79

accounting, and developing a broad range of skills such as adaptability, effective
communication and confidence.
Builds an entrepreneurial mindset: You may use some of the skills of an entrepreneur
in business but even outside of work, it helps in building an entrepreneurial mindset
which creates a positive frame of mind. Even in life, if you want to switch careers
and become an employee, the natural tendency of an entrepreneurial mindset will
set you apart and make you a more favourable candidate.
Develops ability to handle setbacks: Life is full of ups and downs and it is important
to build resilience. The entrepreneurial journey is full of surprises and sometimes
it can be very difficult to accept unfavourable and unexpected negative results.
There are lots of risks involved in every decision you make and a lot of people
judge your skills which can create pressure. A successful entrepreneur will be able
to persevere and handle any failures with ease. They see these setbacks as an
opportunity to learn and grow.
Enhances problem-solving ability: Irrespective of whatever job role you will end up
doing, problem-solving remains the key skill which is very helpful. Even in personal
life, problem-solving is essential. There are lots of conflicts and stressful situations
in the life of an entrepreneur which require the ability to segregate the problem
into smaller components and solve them innovatively. The skill of critical thinking
and analysis is also important while solving complex problems.
Provides exposure to real life: When you take up a corporate job, your role is
specific and focused and by the time you become a CEO, many years roll by.
However, if you start as an entrepreneur, on Day 1 you are the de facto CEO and
need to do everything from scratch with your own hands. The learning curve in
this situation can be extremely steep. This gives an early exposure and knowledge
about the real world.
Creating a positive difference in society: Entrepreneurs fundamentally impact the
economy and society as we read at the start of the chapter. See what Elon Musk
is doing! He created electric cars for a cleaner environment, made space travel
affordable, created digital payments solutions for reducing cost of transactions
and making electronic transactions accessible to everyone.
There are a lot of social impact entrepreneurs who are solving problems of
gender discrimination, inequality, financial inclusion, etc., which help in achieving
the UN SDGs and making the world a better place to live in.

ES.80 Essentials of A rtificial Intelligence- I X

f Writing a note on ‘Entrepreneurship as a career option’: Can you write a simple
note on exploring entrepreneurship as a career option and what would you need
to be ready for being an entrepreneur? Is there an age or skill set to become an
entrepreneur or can anyone start at any time?

\ .... ..... y

«LTHiroiwasi
Collecting success stories of first generation and local entrepreneurs: Do you
have an entrepreneur in the family/relations? Discuss with them how they started
and their business plan. Write down at least three success stories.

L ---- —--------------------------------------------------------------------------------------------------------------J

H5885EP
^Listing entrepreneurial qualities— Analysis of strengths and weaknesses: What do you
think are the key entrepreneurial strengths and weaknesses? Can you list them down?
Remember a strength can also be a weakness, e.g., being optim istic is good but
being too dreamy is bad. Similarly, being hard-working is important but overwork
can stress you out. We will discuss about it in self-management.

^________________________________________________________________________________________

Entrepreneurial S kills- I ES.81

Group Discussion
Have a group discussion on the various skills needed to be a successful entrepreneur
and identify the matching skills that you have:
Skills needed to be a Successful Entrepreneur Skills I Have

How will you develop those qualities that are currently missing in your personality?
Collect information and related data fo r a business: Let us say you are researching
a business idea for using Artificial Intelligence-powered chatbots in e-commerce
companies. Can you identify the sources from which you can collect the below-
given data?
• % of companies globally using Al chatbots
• Growth in chatbot adoption
• Pricing of chatbot solutions
• Industries with the highest use of chatbots
• Benefits of chatbots
Once you’ve identified the sources, please collect and document this data as well.

Meaning of Entrepreneurship Development

Entrepreneurship development is the process of improving the skills and
knowledge of potential entrepreneurs through various learning and development
initiatives. The objective of entrepreneurship development is to increase the
number of entrepreneurs in society and create an ecosystem which promotes
entrepreneurship.
On a societal level, this generates employment and improves the economy of a
country.

ES.82 Essentials of A rtificial Intelligence- I X

ENTREPRENEURSHIP DEVELOPM ENT INITIATIVES
Various entrepreneurship development initiatives have been started to stimulate,
support and sustain the young enterprises. A few of the initiatives adopted by the
Government of India to help the entrepreneurs are (GES 2017):
• Start-up India
• Make in India
• Atal Innovation Mission
• Support to Training and Employment Programme for Women (STEP)
• Jan Dhan - Aadhaar Mobile (JAM)
• Digital India
• Biotechnology Industry Research Assistance Council (BIRAC)
• Stand-up India
• Department of Science and Technology (DST)
• Trade-related Entrepreneurship Assistance and Development (TREAD)
• Pradhan Mantri Kaushal Vikas Yojana (PMKVY)
• National Skill Development Mission
• Science for Equity Empowerment and Development (SEED)

: Objective Type Questions I

1. Which is not a definition of business?

(a) Business is a trade-related activity.
(b) Business involves sale and purchase of goods and services.
(c) Business is done to maximise returns for the shareholders.
(d) Business is done for balance of wealth.
2. Which of these is not a type of business?
(a) Services (b) Manufacturing (c) Social service (d) Hybrid
3. Which of these is not a reward of being an entrepreneur?
(a) Creation of anorganization (b) Improved standard of living
(c) Means of economic development (d) Hard work
4. Which of these is not a role of an entrepreneur?
(a) Coordinating (b) Agent (c) Imitating (d) Capital

Entrepreneurial S kills- I ES.83

 5. Which of these is not the value contribution of entrepreneurs?
(a) Free market evolution (b) New markets
(c) New values (d) Democracy
6. How does society support entrepreneurship?
(a) Provides raw materials (b) Enables financial support
(c) Creates a need for product/service (d) Provides freecapital
7. Entrepreneurship enhances critical thinking and problem-solving ability. (True/False)
8. Entrepreneurship skills can be developed through a structured approach and program.
(True/False)
9. Women entrepreneurship is the process where women organize all factors of
production, undertake risks and provide employment to others. (True/False)
10. Entrepreneurship Development is the process of strengthening the skills and
knowledge of upcoming and existing entrepreneurs by handholding them on their
entrepreneurial journey. (True/False)

: Subjective Type Questions I

1. What are the different types of businesses? Explain with examples.

2. What do you understand by entrepreneurship development?
3. What are the various types of entrepreneurs?
4. What are the roles and rewards of entrepreneurship?
5. What are the core skills of a successful entrepreneur?
6. What roles does an entrepreneur play on a day-to-day basis?
7. What is meant by entrepreneurship management and what are its key components?
8. List down some of the entrepreneurship development initiatives in India.
9. What is the difference between taking up a job and being an entrepreneur?
10. Explain the role of entrepreneurship in a country's economic development.

ES.84 Essentials of A rtificial Intelligence- I X

 o
m ^ *

© ' Green S kills-I

V ®

INTRODUCTION
The term "environment" is used in a broad way. To some people, it simply means
"nature", while for others it means the natural landscape together with all of its
non-human features, characteristics and processes.
These definitions have an underlying assumption that “environment" exists in some
kind of relation to humans. Humans and environment are in association with each
other and need to be in harmony to maintain ecological balance.
atm o sph eric C 0 2

co m b u stio n of
fo ssil fuels

assim ilatio n
by plants and
an im als
anim al
respiration
plant
respiration
assim ilatio n
assim ilatio n by phytoplankton

formation organic decomposition

of fossil fuels (animal waste, dead anim als and plants)
marine deposits of CaCO

f Write a sh
short essay on hazards of deteriorating environment.

G reen S kills -I ES.85

SUSTAINABLE DEVELOPMENT
We will read about UN Sustainable Development Goals (SDGs) in the coming
chapters but it is important to understand that a ‘Green and Clean Planet’ is one
of the focus areas around the world for Sustainable Development.

Relationship between Society and Environment

Society is made up of groups of people living together as part of a community. The
community consumes resources to support the inhabitants. The inhabitants have
their own sets of beliefs and values which affect the environment. For example, if
one group believes that rivers are sacred, then no garbage will be thrown into the
river and they will remain clean. On the other hand, if another group is focused on
consumerism without regard to environment, it may affect the ecological balance.

Environmental
Resource Management
Environmental Protection
Habitat Restoration & Preservation

............................................
A c c o rd in g to C ollins E nglish D ictio n a ry, ‘ E n viro n m e n t co n sists of n atu ral
s u rro u n d in g s in w hich we live o r exist in re la tio n to th e ir physical
c h a ra c te ris tic s ’ .
Every a ctio n has an equal and o p p o site re a ctio n . To a ccu m u la te w ealth
and p ro m o te g ro w th , we e x p lo it resources n atu re has given us, w hile
natu re reacts w ith glo b al w a rm in g , fo re s t fires, c lim a te change and o th e r
m a n -m a d e d isa ste rs. Hence, it m ay be co n clu d e d th a t both so cie ty and
e n v iro n m e n t have a c irc u la r re la tio n s h ip and one im p a c ts the other.

ES.86 Essentials of A rtificial Intelligence- I X

 Human activities
• Industry
r
r
Environment
• Water
l
• Energy production • A ir
•Transport • Soil and mineral
• Agriculture Plants and animals
• Recreation t <r
• Domestic and
personal activities

Wastes/poilutants
.A ir pollutant
• Water pollutant
• Soil pollutant
• Radioactive wastes
• Other toxic minerals

Prepare posters showing environment conservation and share with your classmates.

Write a short essay on various factors that influence our environment.

ECOLOGICAL IMBALANCE
Ecological imbalance is caused when natural or man-made disturbances affect the
natural balance of an ecosystem. Following are some of the man-made disruptions
that cause ecological imbalance: < A U
(i) Deforestation: Due to rising population, we needed to create more cities
and more housing structures which led to deforestation. With increasing
deforestation and climate change, many species of birds and animals are
on the verge of extinction.
(ii) Soil Erosion: Deforestation led to other related problems such as soil erosion
and degrading quality of land making it less fertile.

G reen S kills- I ES.87

 (iii) Overexploitation of Resources: Humans generate tons of waste due to
overuse of resources. Natural resources take millions of years to form and
their overuse can lead to ecological imbalance and cause pollution.
(iv) Environmental Pollution: The pollution caused by industries and vehicles
gets absorbed in land, air and water.
(v) Irregular Land Use: Large-scale mining leads to multiple effects such as land
erosion, dust contaminating the air and improper mining waste disposal.
(vi) E-waste: We all use smartphones and computing devices which generate a
lot of e-waste and need proper management.

NATURAL RESOURCE CONSERVATION

Natural resources such as air, water, land and energy need to be conserved in the
wake of rising pollution. Let us go through the various conservation ideas:
Air Conservation: Minimizing greenhouse effect and reducing air pollutants
such as SPMs will help us conserve fresh air. Using public transport, electric vehicles,
car-pooling, proper maintenance of vehicles, etc., are ways to reduce air pollution.
A / ftw u . 1 Tut,h-t>d
Water Cdnservation: Rainwater Harvesting Systems and waste water treatment
plants are crucial to preserving water for our future generations. Sensor taps help
in reducing water wastage. Aquatic life too needs to be provided adequate fresh
water to conserve it.
Land Conservation: We need to promote afforestation and keep land erosion in
check. This will help in keeping all life forms safe and maintain ecological balance.
Improving soil quality and keeping the waste in check is the key.
Energy Conservation: Conservation of energy means to conserve these resources
and use alternate sources such as wind, solar and tidal energy. New smart cities
are promoting circular energy management techniques to have net zero wastage.

GREEN ECONOMY AND ITS IMPORTANCE

The term “Green Economy" was first coined in 1989 by a pioneering committee in a
report entitled 'Blue Print for a Green Economy’ for the UK government, the committee
consisted of environmental economists D. Pearce, A. Markandya and E.B. Barbier.
According to UNEP (United Nations Environment Program), a green economy
is defined as low carbon, resource-efficient and socially-inclusive economy. The
focus in a green economy is on investments in economic activities that allow
reduced carbon emissions and pollution, improved energy efficiency and general

ES.88 Essentials of A rtificial I ntelligence- I X

sustainability. Investments in these green projects are made through public-private
partnership, policy regulations and taxation reforms.

GREEN ECONOMY =

LOW CARBON RESOURCE SOCIAL

GROWTH EFFICIENCY INCLUSIVE
= SUSTAINABLE DEVELOPMENT?
A Green Economy will help in conserving resources for future generations while
ensuring that more sustainable employment opportunities are generated to
promote economic growth of a country.
U N ECE

• Futures and
Scenarios
• Mining

G o ve rn a n ce
• R e n e w a b le
e n e rg y
• E n e rg y
e fficie n cy
• W ater
e fficie n cy
• Job creation • M o b ility Social justice
Other • Fair trade • In d u s try Income
• In n o va tio n
redistribution

Agriculture

Fig. 5.1: Defining Green Economy

Main Sectors of Green Economy

• E-waste management: E-waste management
means how to handle discarded electronic devices
such as computers, mobile phones, etc. It also
includes ideas for reusing, reselling, recycling or
disposing of such waste.
• Green transportation: Green transportation
includes making efficient and effective use of
resources— types of transport, fuel, pooling of
resources, etc. Green transportation is a key
component of smart cities. It reduces traffic, lowers
dependency on fossil fuels, is cheaper and helps
keep the environment clean.
• Renewable energy: Renewable energy is a form of
energy which is collected from renewable resources
such as sunlight, wind, tide, rain waves, geothermal
heat, etc.
• Green construction: Green construction means to
build sustainable buildings both in terms of structure
and processes that are eco-friendly and efficient
in use of resources. This pertains to the entire
building life cycle process— from design, operation,
maintenance, renovation and demolition.
• Water management: Water management means to
optimize water usage and save it for future use. It
means to have a set of policies and regulations in
place to manage water. Rainwater harvesting is a
popular technique in water management.

* Prepare
Pranara a
anposter showing the importance of green economy with the help of newspaper/
magazine cuttings.

STAKEHOLDERS IN GREEN ECONOMY

Key stakeholders in a green economy are:
• Government
• Environmental Protection and Natural Resources Committee
• Economic Policy Committee
• Agriculture Committee
• Ministry of Environment
• Ministry of Energy and Natural Resources
• Ministry of Natural Resources and Environment

ES.90 Essentials of A rtificial I ntelligence- I X

Role of Government in Green Economy
The government has an important role to play in setting policies that serve, advance
and protect public interest in transitioning towards an energy system that is more
sustainable from both a supply and environmental standpoint. The government
can play a crucial role in this by conducting research and plugging the gaps.

Role of Private Sector

The private sector can collaborate with the government and district administration
in developing smart cities, generating employment opportunities and, thus, creating
a balanced society.
The private sector has resources as well as faster execution capabilities due to
their leaner structure which can contribute to smooth and speedy implementation
of government policies.
The government needs to promote Public-Private infrastructure projects at all levels
so as to build a sustainable governance model.

: Objective Type Questions •

1. Which of these is not a component of Green Economy?

(a) Renewable energy (b) Green buildings
(c) Waste management (d) Fossil fuels
2. Energy that is collected from resources which are naturally replenished on a human
timescale is known as:
(a) Renewable energy (b) Non-renewable energy
(c) Fossil fuels (d) Coal
3. Which of these is not a factor causing ecological imbalance?
(a) Destruction offorests (b) Industrialization
(c) Urbanization (d) Cultural variations
4................... is the term used to denote proper management of a natural resource to
prevent its exploitation, destruction or degradation.
(a) Conservation (b) Reduction (c) Reforestation (d) Green economy
5. Which of these is not a benefit of Green Transport?
(a) Reduces congestion (b) Reduces pollution
(c) Safer and cleaner mode (d) Low mileage and hence costly to use

G reen S kills -I ES.91

 6. Green economy promotes.............a n d .......................
(a) Sustainable development (b) Cost savings
(c) Effective resource utilization (d) Cultural improvement
7. Which of these are the roles of government in a Green Economy?
(a) Supporting research (b) Capital support
(c) Policy-making (d) None of these
8. Why is Water Management important in today's time?
(a) Drinking water is becoming a limited resource.
(b) Floods are increasing water supply.
(c) Rainwater harvesting helps store water.
(d) More people are drinking water now.
9. Green Transport is also known as Smart Transportation. (True/False)
10. Green buildings are also known as sustainable buildings. (True/False)

Subjective Type Questions

1. What do you understand by environment and what is the relationship between
human beings and environment?
2. What do you understand by Green Economy and what are its constituents?
3. What are the main reasons for ecological imbalance?
4. What do you think about the role of technology in creating a balance in the
environment?
5. What are some of the steps that you can take at the community level to curb air
pollution?
6. What is the long-term impact of ecological imbalance in society?
7. What are the benefits of having a green economy and what is its future?
8. What steps can be taken for environment protection?
9. What is the role of government in environment conservation?
10. What do you understand by e-waste management?

ES.92 Essentials of A rtificial Intelligence- I X

 ANSWERS TO OBJECTIVE TYPE QUESTIONS
Chapter-1
1. (d) 2. (b)
3. (d) 4. (d)
5. (d) 6. (c)
7. (b) 8. (b)
9. sender; channel; receiver 10. True

Chapter-2
1. (d) 2. (a)
3. (a), (b) & (c) 4. (d)
5. (d) 6. (d)
7. Stress 8. Time management
9. True 10. True ^

Chapter-3
1. (b) 2. (d)
3. (a), (b) & (c) 4. (a)
5. (d) 6. (a) & (b)
7. (c) 8. Cookies
9. Social media 10. True

Chapter-4
1. (d) 2. (c)
3. (d) 4. (d)
5. (d) 6. (a), (b) & (c)
7. True 8. True
9. True 10. True

Chapter-5
1. (d) 2. (a)
3. (d) 4. (a)
5. (d) 6. (a) & (c)
7. (a), (b) & (c) 8. (a)
9. True 10. True

Green Skills- I ES.93

 o
# * </> *
Inspire and
® 3 A c q u ir e M o dule

Learning Objectives

Ip i To understand and appreciate Artificial Intelligence and describe its applications in

daily life

Ip i To relate, apply and reflect on the human-machine interactions

B a c k g r o u n d a n d C o n t e x t

“The future is already here— it’s just not very evenly distributed.”
— William Gibson

You use Google for searching things but it gives you auto suggestions on what
you might be looking for. You have probably used Siri or Google Assistant on
your phone as well. You shop at Amazon and it gives you recommendations on
purchases based on what you previously bought. You get bored and ask Alexa to
play songs for you. Browse through Amazon Prime or Netflix and it understands
your viewing patterns and gives you recommendations as to what you might like.
Check Facebook or Instagram and it gives you relevant posts and stories on what
you might be interested in. All of these experiences have an underlying technology
at heart— Artificial Intelligence or Al in short.

Inspire and A cquire M odule

Do you realize that you are already experiencing Artificial Intelligence
(AI) and Machine Learning (ML) all around you!

So, what is this Artificial Intelligence anyway?

A rtificial Intelligence is a field of com puter science

aim ed at developing m achines that are intelligent
enough to do certain tasks w hich are otherwise
considered fit to be perform ed by human beings only.
A I uses m achine learn ing to continuously learn and
adapt its algorithm s to becom e sm arter each time and
independently undertakes action without any m andatory
human intervention. M achine learn ing essentially means
m achines learning on their own to im prove their
functioning without hum an intervention.

Artificial Intelligence is by no means a recently discovered technology. In fact, it

was started back in the 1950 s in the US and has been constantly evolving. It has
recently gained prominence owing to the rising internet penetration, proliferation
of connected devices and easy data availability.

The Dartmouth Conference

Stanford researcher John McCarthy coined We think that a significant advance
. the term “Artificial Intelligence” in 1956 during can be made in one or more o f these
1 what is now called the Dartmouth Conference, problems if a carefully selected group /
_ in which the core mission of AI was defined. o f scientists works on it together for a
In the original proposal for the conference, summer.
McCarthy framed the effort with the following: —John McCarthy, Marvin
An attempt will be made to find how Minsky, Nathan Rochester, and
to make machines use language, form Claude Shannon, “A Proposal for
abstractions and concepts, solve the Dartmouth Summer
kinds o f problems now reserved for Research Project on Artificial
humans, and improve themselves. Intelligence" (1955)

Fig. 1.1 : The term AI was first coined during the 1956 Dartmouth Conference

2 E ssen tia ls of A r t if ic ia l In t e l l ig e n c e - I X
IMPORTANCE OF ARTIFICIAL INTELLIGENCE
According to Google CEO Sundar Pichai, Al will have a bigger impact on humanity
than some of the other well-known innovations.

“A I is one o f the most im portant things hum anity is w ork in g on. It is

more profound than electricity or fire.”
—Sundar Pichai

Al has managed to disrupt the conventional functioning of most of the work sectors
and continues to expand its influence over the remaining ones. The most interesting
aspect of Al is that it enables algorithms to handle situations and solve problems
on their own without human supervision, and learn and improve over time using
Machine Learning. Sounds amazing, isn’t it?
are num erous possibilities o f A I in our daily life. From (#
self-d riv in g cars in the US to traffic m anagem ent robots in Dubai,
A I is all set to revolutionize the w o rld in w ay s w e just can’t even
im agine. According to the W all Street Journal, quoting an industry
source, more than 50,000 job opportunities in data science, A rtificial
Intelligence and M achine L earn in g are still vacant due to shortage o f
qualified trained m anpower, w hile globally, alm ost 40% jobs in A I are
lying vacant for lack o f trained people. The average sa la ry o f a
Data Scientist w ith M achine Learn ing skills is ? 874,691 per year.
(So urce : PayScaie)

The future will not be about AI vs non-AI but people who know AI vs
those who don’t, and guess who will win!

■Summarize a list of daily activities that you do which you believe m ight involve
Artificial Intelligence.

~ ......

* Make a list of the devices that you use which m ight involve Al technology. Also write
the manner of Al involvement.
v___________________________________________________________________________V

I nspire and A cquire M odule 3

 ____ DATA NEVER SLEEPS 6.0
^ J® , How much data is generated e v e ry m in u te ?
There's no way around it: big data just keeps getting bigger. The numbers are staggering, but they’re not
slowing down. By 2020, it's estimated that for every person on earth. 1.7 MB of data will be created every
second. In our 6th edition of Data Never Sleeps, we once again take a look at how much data is being created all
around us every single minute of the day—and we have a feeling things are just getting started.

■m
NETFUX m U
N
■:•••lam GAINS

NEW " — "

(RECAST PROFESSIONALS

MJUUN ~ REQUESTS

” 1,111 USERS WATCH

VIDEOS

USERS PUBLISH USERS SEND

2018
L T L 'I'U 473,400
12,986,111
TEXTS SEN T

w w ir n /m m
r USERS MATCH / PROCESSrt ise ?S TAM: \ s’ ' lA M ic v t' \ — USE—

6,940 7 s68'493/1,389 \750,000X3,138,420

The w o rld 's in te r n e t p o p u la tio n is g ro w in g
s ig n ific a n tly ye a r-o v e r-y e a r. In 20 1 7 , in te rn e t T h e a b ility to m ake d a ta -d riv e n d e c is io n s is
usage re a c h e d 4 7 % o f th e w o rld s p o p u la tio n c ru c ia l to a ny business. W ith e ach click,
and n o w re p re s e n ts 3.8 b illio n p e o p le . s w ip e , share, a n d like , a w o rld o f v a lu a b le
in fo rm a tio n is c re a te d . D o m o p u ts th e
p o w e r t o m a ke th o s e d e c is io n s rig h t in to th e
p a lm o f y o u r h a n d b y c o n n e c tin g y o u r d a ta
a n d y o u r p e o p le a t an y m o m e n t, o n any
d e vice , so th e y can m a ke th e k in d o f
d e c is io n s th a t m a ke an im p a c t.

L e a rn m o r e a t d o m o .c o m
GLOBAL INTERNET POPULATION GROWTH 2012-2017
(IN B IL L IO N S )

SOURCES: STATISTA, LINKEDIN, INTERNET LIVE STATS, EXPANDED RAMBLINGS, SLASH FILM, RIAA, BUSINESS OF APPS,
INTERNATIONAL TELECOMMUNICATIONS UNION, INTERNATIONAL DATA CORPORATION

Fig. 1.2 : Data produced on different internet platforms

From the above chart, you can see the sheer amount of data and content we are
generating per minute on the internet. It is amazing to see all this data getting
built up on the internet. Imagine the possibilities of using this data for analytics
and personalization of user experience.

4 Essentials of A rtificial I ntelligence- I X

 Artificial Intelligence Timeline

Fig. 1.3: History of Al

This chart shows you how Al has evolved over the years and how it is not a new
technology at all. From science fiction in the 1940s, when Al first captured human
imagination, to the 1956 Dartmouth Conference and then Turing Test, to chatbots
and IBM Watson now, Al has taken over six decades of research to reach where
it is today.

INDUSTRY 1.0
M e c h a n iz a t io n , s t e a m M a s s p r o d u c t io n , A u tom ation, c o m p u te rs Cyber Physical System s,
p o w e r , w e a v in g loom a s s e m b ly lin e , e le c t r ic a l and e lectronics internet of th ings,

Fig. 1.4: Evolution of Industry (Source: ic t N ew s)

Inspire and A cquire Module 5

WHY AI IS GAINING RELEVANCE
As we have seen in Fig. 1 .1, AI is by no means a recent discovery and has been in
existence for over 50 years which makes one wonder why it has been making so
much noise of late. The key reasons for its recent popularity are:
<§ Increased and more than ever powerful computational resources which enable
the heavy AI algorithms to work. Do you know that an average smartphone has
more computational power now than was used on the entire Apollo mission
which got man on the moon!
<§> AI and machine learning needs training data to be able to get better with time.
More the sampling data, more improved the outcomes. We now have so much
data available with billions of people using cheaper smartphones with so many
apps and constantly generating more personal data (Fig. 1 .2).
<§> Internet and Data plans have become cheaper with increased internet
penetration, thus making more and more people and devices get connected
and enable AI globally.
<i) We now have access to apps such as Siri, Alexa and Google Assistant that know
so much about us— our preferences, location, profile, etc., and offer us relevant
content and information, enabling a shift from Narrow AI to Broad AI.

Broad AI: Broad AI systems are capable of executing various tasks across
various fields. Imagine a robot which can do your laundry, understand
your voice commands for reading emails, managing calls and scheduling
appointments—all at once. Broad AI would truly replicate human
intelligence and help us leverage true power of AI.
Narrow AI: Narrow AI systems are very good at one specific task that they
are designed to do. They can’t execute any task outside their scope. Imagine
an image recognition system that is designed to distinguish between
humans and animals but cannot tell the difference between dog and cat
unless it is so designed.

Do you know that around 2.5 quintillion bytes of data is produced

every day, preparing a solid foundation for large AI applications!

It is estimated that in 2020, 1.7MB of data was created every second for
every person on the Earth.

6 Essentials of A rtificial Intelligence- I X

 Can you identify examples of Narrow Al and Broad Al from your daily life applications?

Types of Al

“As soon as it works, no one calls it Al anymore.” — John McCarthy

As we have seen, with the rise in the number of devices, apps and connectivity, we
are moving more and more towards a Generic Al or General Al which can handle
multiple tasks at the same time.

“ I find it useful to distinguish what I will call “strong” Al from “weak” or

“cautious” Al. According to weak Al, the principle value of the computer in
the study of the mind is that it gives us a very powerful tool. For example, it
enables us to formulate and test hypothesis in a more rigorous and precise
fashion. But, according to strong Al, the computer is not merely a tool in
the study of the mind; rather, the appropriately programmed computer
really is a mind, in the sense that computers, given the right programs, can
be literally said to understand and have other cognitive states.”
— Searle, 1 9 8 0 , M inds, B rains a n d P rogram s
Wm w P __ PWPPWfP

What this essentially means is that as we move towards a stronger Al capability,

we will start to look at mirroring the way our mind works, thinking of many things
at the same time rather than only focusing on one task. Currently, all Al is largely
task-driven where a specific activity can be done by the machine.
When Al works around specific tasks only in closed domain, it is called Narrow Al
such as Asking Alexa (Amazon home automation system) to play a song or tell a
joke from a library of jokes and songs connected through internet (Amazon Skills,
Amazon Music, etc.). However, when Al works in general, such as with IBM's
Watson (IBM's Al system), it is called Generic Al. Most examples around us in our
lives today are of narrow Al.

There is also merit in having a few independent narrow Al systems and then
bundling them into a larger broad Al system. A perfect example of it is a driverless
car. There are so many sub-systems at play such as (indicative, not exhaustive):
• Image recognition (seeing objects around the car)
• Speech recognition (voice commands from user)

Inspire and A cquire Module 7

 Central intelligence console that prevents crash and plans trajectory/lane
driving (using sensors for object detection around the car)

Wireless Communication
Paik Assistance
Sufround View

R « r Collision
Warning , Surround View

Long-Range Radar

Camera LIDAR
Ultra Sound"
(Source: h ttp s://w w w . la n d m a rk d iv id e n d . c o m )
Fig. 1. 5: H o w a d r iv e r le s s c a r w o r k s

To figure out how a self-driving car works,

Scan QR code or visit:
https://www.yoL/tube.com/watch ?v=taMP_n3wL 7M)

Alexa: Amazon Alexa, simply known as Alexa, a virtual assistant developed

by Amazon, was first used in the Amazon Echo and Amazon Echo Dot smart
speakers developed by Amazon Lab 126. It is capable of voice interaction,
music playback, making to-do lists, setting alarms, streaming podcasts,
playing audiobooks, and providing weather, traffic, sports and other real­
time information, such as news. Alexa can also control several smart
devices using itself as a home automation system.
IBM's Watson: Watson is a computer system capable of answering
questions posed in natural language, developed in IBM’s DeepQA project by
a research team led by principal investigator David Ferrucci. Watson was
named after IBM’s founder and first CEO, industrialist Thomas J. Watson.
• LIDAR: Light Detection and Ranging is a remote-sensing method that
uses light in the form of a pulsed laser to measure ranges to the Earth.
(Source: h t tp s :/ / tc f.p a g e s .tc n j.e d u / file s / 2 0 1 7 / 0 5 / A u to m o tiv e -S e n s o r s -JJe s s o n -V 3 .p d f)

• Long-Range Radar: It is a premium long-range radar sensor for forward-

looking functions specially made for Highly Automated Driving.

Another perfect example of such a broad Al system could be smart home systems.
 Do you know that Bill Gates’ house in Washington has many interesting
gadgets to optimize energy consumption and enhance visitor experience.
A small example is that when a guest arrives, they are given a pin
(a simple gadget that can be held or put on their clothes) that interacts
with all different sensors in the house and the guest can enter their
temperature and lighting preferences which change as they move from
room to room. Music also keeps following you as you navigate the house
and adapts to temperament. Even the artworks on the walls can
be changed through touch sensors. Sounds amazing! Doesn’t it?

Rough layout of the floor plan of a dream smart home

Think about how your dream smart home would look like? What would be its features
and amenities? What kind of gadgets and sensors would it have? Imagine looking
at the home from above. How does it look? Draw it.
V____________________________________ _____________________________________ y

EVOLUTION OF AI
Artificial Intelligence (AI) Trends
Independent of Application
N a rro w A I - AI designed to execute a single task better than a ???
human
U b e r B o t - Bots capable of answering any structured query in
conversational style of a human Superintelligence
E n h a n c e d N a rro w A I - Narrow AI capable of more than one task
N a rro w A I C l u s t e r - An assem bly of Narrow AI system s to
achieve greater capabilities
N a rro w A I S u p e r c l u s t e r - An assem bly of Narrow AI Clusters ' General AI Acceleration
C G e n e ra l A I - AI with full cognitive capabilities mimicking a human (with or without
<O
c S u p e r in te llig e n c e - AI well beyond human cognitive capabilities
consciousness)
Possible
<
Q_

o
........................................................................ ' - ........................................... ............................... M |i

Narrow AI
/
Superclusters
< *
00
Narrow AI
Clusters Convergences Driving Acceleration
oo .c f • Internet of Things
Enhanced • Next-Gen Chips (Moore’s Law 2.0)
Narrow AI • Supercomputers
• Sensors
UberBots • Robotics
(“Al-first”) • Cloud
Narrow AI
Deep Learning (imagery, language,
etc.) Bots (conversational, search, etc.)
2016 2017- 2018 ???
2018 FutureGrasp, LLC
Fig. 1.6 : Evolution of Artificial Intelligence

I nspire and A cquire M odule 9

As you can see from the above graph, in recent years, the focus of Al research has
been on conversational chatbots (covered in detail in the next section) and other
narrow Al systems. But with the advent of other technologies such as Internet of
Things (loT), new age fast chips and sensors, it is becoming easy to build much
more complex and smarter Al systems. Such semi-complex systems are known
as clusters.

The focus of future research is increasingly on building machines that can truly
think like humans as well as coordinate and collaborate with them. This would be
termed as General Al wherein machines can autonomously think and create new
solutions for unknown situations or scenarios, just like humans!

“Al can have two purposes. One is to use the power of computers to augment
human thinking, just as we use motors to augment human or horsepower.
Robotics and expert systems are major branches of that. The other is to
use a computer’s artificial intelligence to understand how humans think. In
a humanoid way. If you test your programs not merely by what they can
accomplish, but how they accomplish it, then you’re really doing cognitive M
science; you’re using Al to understand the human mind.” !■
-Herb Simon ,
Humans have always tried to see how machines can think more and more like us
as our brain is fascinating in the way it works. Earlier, we only thought of machines
to help us with labour-intensive work or hard tasks, but increasingly with Al, we
are looking at machines to truly replicate our thinking process. In such a world,
humans and machines could really collaborate with each other.

Currently
Technology ■*
is here

Fig. 1.7: Stages Of Al (Source: w w w .x e n o n s ta c k .c o m )

10 E ssentials of A rtificial Intelligence - I X

 Humanoid: A robot which looks like a human in its appearance and
conversations.
Internet of Things: Internet of Things (IoT) is a system of interrelated
computing devices and sensors that are provided with unique identifiers
and the ability to transfer data over a network without requiring human-
to-human or human-to-computer interaction.
Scan QR code or visit:
http$://matscloud.blogspat.com/2017/02/understanding-iot~protocols-mqtt-caap;html

• Neural Network: A neural network is a series of algorithms that aims to

recognize any underlying relationships in data set through a process that
mimics the way the human brain operates.
• Deep Learning: Deep learning is a subset of machine learning in artificial
intelligence (AI) that has neural networks capable of unsupervised learning
from data that is unstructured.
• Moore’s Law: Moore’s Law is the observation made in 1965 by Gordon
Moore, co-founder of Intel, that the number of transistors per square inch
on integrated circuits had doubled every year since the integrated circuit
was invented.

WHAT IS INTELLIGENCE IN AI—THE TURING TEST

When AI started gaining prominence, there was a constant debate as to how do
we know that a system is intelligent. Since humans have always believed to be
at the top of the food chain, we assumed that any machine that could think like
humans must be intelligent.
In the 1950 s, Alan Turing proposed that a human evaluator will judge the
conversations between humans and machines. The conversation channel would
be text only and no speech would be used. If the evaluator can’t tell the machine
from human, then the machine is said to have passed the Turing Test.
At its core, the Turing Test uses:
• Natural Language Processing (NLP): It is a subfield of artificial intelligence which
helps design systems on how to process and analyze large amounts of natural
language data. For example, if you are talking to Alexa, it needs to understand
your language, words, context and emotion as well. NLP is the technology
which enables Alexa to accomplish this task.

Inspire and A cquire Module 11

• Knowledge Representation: It means to organize and represent knowledge
in a form that systems can access and learn intelligently from that knowledge
database.

• Automated Reasoning: It means applying reasoning in the form of logic to

computing systems. If we are given a set of assumptions and a goal, an automated
reasoning system should ideally be capable of making logical inferences towards
that goal automatically.

• Machine Learning (ML): As already explained, it means when machines are

able to intelligently learn on their own to improve their functioning without
human intervention.

• See how Google Duplex came close to passing the Turing Test

Scan QR code or visit:

https://youtu.be/D5VN56jQMWM

EXPERIENCING AI IN REAL LIFE— CHATBOTS

In 1966, ELIZA was the first natural language programming (NLP)-based conversation
program described by Joseph Weizenbaum.

It featured a conversation between a human user and a computer program

representing a mock psychotherapist.

The original program was implemented on IBM 7094 of Project MAC time-sharing
system at MIT and was written in MAD-SLIP AI language.

MAD: Michigan Algorithm Decoder (MAD) is a programming language

and compiler for IBM 704 and later for IBM 709, IBM 7090, IBM 7040,
UNIVAC 1107, UNIVAC 1108, Philco 210-211, and eventually for IBM
S/370 mainframe computers.
SLIP: Symmetric List Processor was first implemented as an extension to
the Fortran programming language, and later embedded into MAD and
ALGOL.

12 Essentials of A rtificial Intelligence- I X

 EEEEEEEE 1 IIIIIII ZZZZZZZ 000
E 1 I z 0 A
E L I z 0 A
EEEEE L. I z 0 A
E L I z 0AA0000
E L I z 0 A
EEEEEEEE ilium iiiiiii im n A A

ELIZA > (PLEASE TVPE IN ALL CAPS} WHAT'S VOUR NAME DEAR ?
AMTT
ELIZA > HI. I ’M ELIZA. WHAT DO VGA WANT 10 TALK ABOUT ?
AMI I > CON I TALK ABOUT MV PROBLEM ?
ELIZA > SURE... '
OMIT > DUE TO ON GOING EXAMS IN COLLEGE, I MOVE 0 LOI OF STRESS.
ELIZA > PLEASE RELAX 0 LITTLE. SLEEP WELL.
OM.LI > 1HONKS LOR VOUR ADVICE.
ELIZA > NO MFNTIQN.
OMIT > BVE
ELIZA > BVE AND KEEP IN TOUCH...

Fig. 1. 8: Actual screenshot Of Eliza chatbot (Source: E liza P r o je c t)

For a little chat with Eliza from 1977,

Scan QR code or visit:
https:/./web. njit. edu/~ronko wit/eliza.html

Fast forward to 2019. We are now surrounded by chatbots or bots in general— from
Alexa to Siri to Google Assistant.
Chatbots are essentially assistants that communicate with humans through text
or voice. They are virtual companions that seamlessly integrate into websites,
applications and even instant messengers, and help businesses get closer to their
users.

Depending on how they are programmed, they can be categorized as:

• Simple Chatbots: They work on pre-written keywords that they understand.

Each of the commands that they are going to follow must be coded into them
by the developer. So, if a user asks them something outside of their knowledge
base, they respond with “ Sorry, I did not understand” , or something along these
lines.
• Smart Chatbots: Smart chatbots are in trend today! Based on Al, these bots
don't have pre-programmed answers. They learn with time, catching keywords
and putting them in context, and help users arrive at the most relevant answers
to their queries.

Chatbots are something every business should consider. Apart from just getting
rid of routine tasks and processing multiple queries at once, chatbots also go a
long way in gaining customer loyalty.

Inspire and A cquire Module 13

The reasons why customers prefer chatbots are as follows:

• Productivity: Chatbots provide fast and efficient access to information thereby

improving productivity.

• Entertainment: Since chatbots are Al-powered, they can hold amusing

conversations with people while they are waiting on a platform. This also helps
keep users engaged to the site.

• Social and Relational Factors: Fuelling conversations, chatbots enhance

social experiences of visitors. Not only do chatbots help users speak up
without being judged, they also help them enhance their conversational
skills.

• Curiosity: People are always curious about trying something new. And the
novelty of chatbots sparks amazing curiosity.

With all those benefits and more, users find it amazing to get assisted by a chatbot.
There is no denying that as these bots get more sophisticated, more and more
businesses will embrace them.

Do you know that Sophia was the first robot that was
recently offered citizenship by the Kingdom of Saudi Arabia.

. ..... chatbots
Find out chatbots on the internet and interact with them to understand and appreciate
how they use NLP to have an engaging conversation with the users.
J

14 E ssentials of A rtificial Intelligence - I X

IBM WATSON— A BIG LEAP IN AI

For Watson and the Jeopardy! Challenge,

Scan QR code or visit:
E- https:/.'/www..youtube, com/watch ?v=P 18EdAKuC1 U

For most people globally, their first encounter with AI was with IBM Watson—
the computer system developed to answer questions on the famous quiz show
Jeopardy!* (a US-based reality show) and, in 2011, the Watson computer system
competed on Jeopardy! against legendary champions Brad Rutter and Ken Jennings,
winning the first place prize of $1 million!

Underlying Watson was 16 terabytes of RAM through which it could process 500
gigabytes (1 million books) per second. Content was stored in Watson’s RAM rather
than in memory for easy access and it cost about $3 million.

How Watson Works

• It receives questions or inquiries as electronic texts.

• Divides these texts into different keywords and fragments and searches for the
intent from texts (signal classification from noise).

• Quickly executes thousands of language processors/algorithms.

• As more and more algorithms look for the same answer, it boosts Watson’s
confidence and helps it to make the right response/guess.

• The show features a quiz competition in which contestants are presented with general knowledge
clues in the form of answers, and they are required to phrase their responses in the form of questions.

Inspire and A cquire Module 15

AI IN INDIA

Prime Minister Narendra Modi said in 2018 that the evolution of technology has
to be rooted in the ethics of “Sab ka saath, Sab ka vikas”. (We need to make AI
in India and it should work for India.) He also said that IT + IT = IT, which means
that India's Talent + Information Technology = India Tomorrow and AI sits at the
heart of this national technology strategy.

NITI Aayog set up a #AlforAII hashtag and came up with a discussion paper on
National AI Strategy. In a country as diverse as ours with limited infrastructural
capabilities and suboptimal efficiencies, AI could be a real boon in improving access
to and quality of all e-governance services.

Imagine if AI could be used in voting, we would have 100%voting with everyone

sitting at home and voting and AI monitoring it. Or if there were Al-enabled police
eliminating the need to report any crime as it would pre-empt a crime-like situation
and intervene early. You have to think in such a revolutionary (and not merely
incremental and evolutionary) way when you have such an amazing technology
at your disposal.

FUTURE OF AI

AI as a technology is evolving with each passing day and we are seeing newer
applications in all sectors such as health care, education, banking, logistics and
retail among others.

In future, we will see a huge AI impact on automated transports. Humans

are generally fragile and tend to make mistakes, especially when it comes to

16 E ssentials of A rtificial Intelligence - I X

transportation and driving. Automated transportation will ensure there are fewer
accidents. This will also allow the transportation systems to flourish and help the
industry make huge profits.

Another way Al will evolve is its use in conjunction with our brains. Augmentation
is a hot topic and its future looks promising, especially when coupled with Artificial
Intelligence. Researches are extremely positive when it comes to Al augmentation
in the future.

Hazardous jobs, too, will be a thing of the past— well, for the most part. Al will be
able to handle many of them as it doesn’t understand the concept of pain and can
be reassembled if destroyed. This opens a lot of possibilities such as bomb defusing.

Al can also be used to overcome several problems associated with our environment
today. Al, in congruence with big data and other technologies, can help solve
problems relating to climate change, natural calamities prediction, etc. Even though
there is a fair bit of time before we reach that stage, the possibilities are endless.

As Al evolves and becomes smarter, the sky will be the limit. All we know for sure
is that Al will overhaul every possible domain’s operations in times to come.

FUTURE WORK SKILLS WITH Al

Since we are in the midst of the Fourth Industrial Revolution, it is time to start
preparing for future work skills. With Al coming into being, the nature of skills
requirements will also change and evolve. With fast-paced technological innovations,
most of us would soon be sharing our workplaces with artificial intelligence tools
and bots. This means machines will be able to do our work better, faster with
more efficiency. Humans and machines will be able to complement each other
and machines may even go beyond human limits.

Now, the apprehension that the future will witness huge layoffs or
that there will be a workforce sans humans is being unrealistic. It is
true that some occupations will lose steam, but then there are
others that will grow, and many more will emerge. So,
how can we be future ready?

For this, it will be crucial to trace the pattern of job

change which can be easily done by identifying the
required skill set for different jobs. Most likely, half
of the jobs currently done by humans, especially those
carried out in a very structured and organized manner
or those requiring physical and manual skills, will be
automated. However, those requiring managerial skills, higher
cognitive skills or emotional and interpersonal skills are likely to witness a surge.

Therefore, as a learner, it is important to acquire the following skills to stay relevant

in a future workplace.

1. Creativity: We all will need to be creative enough to work hand in hand with
technology to fully tap the potential benefits. Al is not intended to compete
with humans; rather we have to design better systems to take on higher order
cognitive tasks. The future workplace will have both humans and machines
working together.
2. Emotional Intelligence Quotient (EQ): Emotional Intelligence means having
empathy towards others, high integrity and collaborating with others. Machines
cannot replace our ability to connect with other people and that is why people
with higher EQ will be in demand. Therefore, jobs requiring emotional intelligence,
like nursing, will be more relevant in future.
3. Analytical Thinking: A person who can come up with innovative solutions and ideas,
solve complex problems using reasoning, evaluate arguments and analyze the flow
of information from various resources while keeping in mind the constraints is
least likely to be overpowered by Al and automation. People with strong analytical
thinking will be needed to navigate the human/machine division of labour.
4 . Active Learning: One needs to be on the lookout for new learning needs and be
adept at acquiring these skills. Gone are the days of passively reading books and

18 E ssentials of A rtificial Intelligence - I X

10. Embrace Constant Change: Workplaces of tomorrow will change very quickly
and constantly evolve with more and more human-machine interactions.
Therefore, it is only right to see change as an opportunity to grow rather than
be overwhelmed by it.
The nature of work, skills required and job division between humans and
machines will dynamically change so often in coming years that one has to be
ready and flexible to adapt quickly.
Given that the half-life of a skill has dropped from 30 years to an average of 6
years, it is important to become a lifelong learner and keep acquiring new skills
every few years. One has to be future ready at all times and invest in learning
today for what maybe relevant in the near future. So, what’s the most important
skill you should acquire starting today?
Al is the answer!

; Objective Type Questions )

A. Fill in the blanks:

1 .................... was the first chatbot invented in 1966.
2 ...........................is the first humanoid to be granted Saudi Arabian citizenship.
3. NITI Aayog set u p .......................... mission to see how Al can be used in all aspects
of a citizen's life.
4. Voice-based command systems u se......................... to understand and interpret the
commands.
5. When machines can think and create new solutions for unknown situations on their
own, it will be called............................
6. Most chatbots u se..........................to answer user queries.
7. Most Al programs u se ......................... as the computer language.
8. Machine Learning is a ..........................of general field of Artificial Intelligence.

B. State whether the following statements are True or False.

1. Google Assistant has passed the Turing Test.
2. Facial recognition-based unlocking of a phone is called broad Al.
3. Alexa is an example of Al.
4. The process of analyzing data to derive patterns from it is called Analytics.
5. The 4th Industrial Revolution is led by advancement in telecom.

20 E ssentials of A rtificial Intelligence - I X

 6. Emotional Intelligence Quotient (EQ) is a key future world skill.
7. General Al will have consciousness.
8. Humanoids are robots with human interface and capabilities.

C. Multiple Choice Questions (MCQs):

1. A Professor at the Stanford University coined the term "Artificial Intelligence" in 1956
at a conference held at Dartmouth College. Name the Professor.
(a) AlanTuring (b) John McCarthy (c) Andrew Ng (d) Gary Marcus
2. Which was the first chatbot ever made?
(a) Eliza (b) Alexa (c) Siri (d) Google Assistant
3. Name the robot that beat Garry Kasparov in chess:
(a) Siri (b) Watson (c) DeepBlue (d) Sophia
4. Which is not anexample of ImageRecognition Al?
(a) Seeing an image to identify the object in it
(b) Identifying humans or animals in an image
(c) Understanding emotion of a human in the image
(d) Adding a colour effect to the image
5. What was the reality game show in which IBM Watson beat the existing champions?
(a) The Price is Right (b) Jeopardy!
(c) Who Wants to Be a Millionaire? (d) Wheel of Fortune
6. Which is the famous human intelligence test?
(a) Watson Test (b) Turing Test (c) Deep BlueTest (d) Andrew Test
7. What is the technology that helps machines understand and interpret a text sentence
to provide relevant response?
(a) Natural Language Processing (NLP) (b) Internet of Things (loT)
(c) Blockchain (d) Augmented Reality
8. What is an example of complex Al system comprising mini Al sub-systems and enabling
technologies?
(a) TV remote (b) Self-driving Car (c) Chatbots (d) Printer
9. How much data would be generated per person per second in 2020?
(a) 10MB (b) 1.7MB (c) 1GB (d) 2.5KB
10. Roughly how many books per minute can IBM Watson process?
(a) 100,000 (b) One million (c) Two million (d) 500,000

Subjective Type Questions :

1. Can machines ever surpass human intelligence? Explain your thoughts in support of
your argument.
2. Is Al a recent technology? If no, explain why it has gained prominence only recently.
3. Do you think Al is good for humanity or do you foresee problems with the technology?

Inspire and A cquire Module 21

 4. How do you think Al is already impacting your present-day life? Give examples.
5. What do you think can happen if Al falls in wrong hands? What kind of challenges
could Al pose to society?
• 6. What are the typical benefits of chatbots for a business?
7. What new-age technologies in Al do you think can surpass the Turing Test?
8. Do you think Al will result in more and more jobs going to machines? What will
humans do if machines do all the work?
« 9. What could be the applications of Artificial Intelligence in governance in India?
10. List some use cases of Al in public services such as passport office, electricity, water,
waste management, etc.

__________________

Quick Activity
j ____ i____ i____ i____ i____ i____ i____ i____ i____ i____ i____ i____ i____ i____ i____ i____ i____ i

Think and plan a small presentation on how Al could be used in classrooms. Which subjects could be taught
through Al and how would teaching be different in an Al-enabled classroom?

• Play the Emoji Scavenger hunt and share your observations.

Scan QR code or visit:

https://emojiscavengerhunt.withgoogle.com

• Play Rock, Paper, Scissors and list down how you think it works and what are the underlying principles.•

Scan QR code or visit:

http://rockpaperscissors.ai

• Try playing with QuickDraw. See how fast it predicts after trying a few times. Understand how it works
and discuss with your peers.

Scan QR code or visit:

https://quickdraw. withgoogle, com

22 E ssentials of A rtificial Intelligence - I X

 A
- :/> *

D omains of Al
W (D ata, CV and NLP)

Learning Objectives
Ip | Understand and identify the three key domains of Al
Ip | Learn to apply these domains in Al-led problem-solving

B ackg ro und a n d C o ntext

“ Some people call this artificial intelligence, but the reality is this technology will enhance
us. So instead of artificial intelligence, I think we’ll augment our intelligence.”
— Ginni Rometty

Just like any other technology, Al in its true sense will support humans in doing
their work better. The positive side of Al is that it can support humans in higher
cognitive abilities-led tasks as well which was not possible earlier.

It is interesting to learn how Al is able to carry out such complex tasks and what
are its underlying capabilities. Al has three underlying domains: Data, Computer
Vision (CV) and Natural Language Processing (NLP). We will now look into details
of each of these and then interpret them together in action as well.

DATA IDENTIFICATION AND COLLECTION

We saw in Chapter 1, Fig. 1.2 that as a society, we are generating data at an
unprecedented rate. These data sets can be numeric (sales, insurance premium,
weather data, etc.), categorical (colour, gender, etc.), or even unstructured free
text (comments, notes, feedback).

Data collection is the process of identifying various sources of data (structured

and unstructured data), collecting the data and preparing to label it. We need
to make sure that data collected is in the correct form at and aligned with the

D omains of Al 23
project requirements. For example, if you are to make a weather prediction model
using Al, you need to identify what parameters define weather such as temperature,
humidity, etc., and all the underlying patterns that impact weather. We need to
then define the sensors from which this data can be collected and so on.

The key to data collection is to structure the data in such a format that one can run
analytics and pattern recognition models on top of this data. From those patterns,
one can build predictive models using machine learning algorithms that look for
trends and predict the future.

BID
Al-Enabled Duet
Scan QR code or visit:
https://experiments. with google, com/ai-duet

This experiment lets you play a duet with the computer. Just play some notes and
the computer will respond to your melody. You don’t even have to know how to play
the piano— it’s fun to just press some keys and listen to what comes back. You can
click on the keyboard, use your computer keys, or even plug in a MIDI keyboard.
It’s just one example of how machine learning can inspire people to be creative.

Predictive and forecasting models are completely dependent on the quality of

input data and its consistency. “ Garbage In, Garbage Out (GIGO)” holds true in

24 E ssentials of A rtificial I ntelligence - I X

this case. So, focus on ensuring that you get the right data set, clean it and check for
consistency and hygiene. In many Al projects, data sets are sometimes incompatible
with the output intended.

For example, in the above instance, if you are to predict weather data for a city
(say Mumbai, which is coastal), getting data sets for a different city whose weather
is very different from the city in question (say Udaipur, which has a semi-arid
climate) will not serve much purpose.

(Source: ‘The A l H ie r a r c h y o f N e e d s ' b y M o n ica R o g a t i)

Fig. 2.1 : Monica Rogati’s Data Science Hierarchy of Needs Pyramid

Similar to Maslow’s Hierarchy of Needs, Monica Rogati’s Data Science Hierarchy of

Needs is a pyramid framework to understand intelligence levels in an organization.

It starts with having basic hardware, sensors and devices in places to capture the
data required for our Al model. We need to ensure right data, right formats and
quality of data at this stage.

Next is having the right storage systems which means having IT infrastructure
(servers, cloud storage, etc.) and systems such as ERP, MIS, which can keep
this data.

We then need to do data cleaning to ensure that right data in correct format
(structured data) is available to run any analytics or data science models on it.

D omains of Al 25
The next step is to run data visualization models, classification of data, data
labelling and defining some analytics metrics for the identified set of data.

The final stage is to apply Machine Learning algorithms to identify patterns

and forecast future trends on data. We need to do A/B testing to iterate on the
model.

The interesting thing about creating data sets in your respective industry
is that you can use it for competitive advantage and create solutions around
it. Companies like Google, Amazon and Facebook are dominating their
industries because they were the first to begin building data sets. Their data
sets have become so large and complicated and their data collection and
analysis is so sophisticated that they are able to grow it to their competitive
advantage.

Adopting Al and ML is a journey and not a magic wand to produce results instantly.
It begins with gathering data into statistical processes and simple visualizations
that allow you to better interpret your data and get your processes under control.
From there, you,ll progress through increasingly advanced analytical capabilities
until you achieve that utopian goal of self-learning algorithms that optimize your
process on their own without human intervention.

26 E ssentials of A rtificial Intelligence - I X

 Things to Remember r..........................................
We have a lot of data available now in the form of rich media such as pictures, video
and voice files. In the past, we used to minimize this type of data that we captured
because we couldn’t do much with it as it required large computational abilities and
involved high storage costs.

BIG DATA AND AI

There is a mutual reciprocal relationship between Big Data and AI.
AI heavily depends on Big Data structures for success while at the same time
helping organizations unlock the potential of their data warehouses in ways that
were earlier very difficult or impossible to achieve.
The focus today is on generating more and more data from a variety of sources so
as to improve the machine learning models. We can now process both structured
and unstructured data.
Historically, when it came to analyzing data, the only option was to use a query
or SQL (a long list of complicated queries). Now with Big Data, we have newer
methods of pre-processing and analyzing data.

AI is the next logical step to SQL-based search. Organizations need to combine the
power of human intuition with machine intelligence to augment these technologies
— to create what is called Augmented Intelligence. More specifically, an AI system
needs to learn from data as well as from humans in order to be able to fulfil its
function.

AI is becoming a circular and ongoing process with Big Data. First, data needs to be
into the AI engine, making AI smarter with underlying machine learning algorithms.
With smarter AI, lesser human intervention is needed for the AI to run properly.
And finally, the less AI needs outside involvement to run, the closer the society
comes to realizing the full potential of this ongoing Al/Big Data cycle.
(Source: M a ry v ille U n iv e rs ity )

You can use the following data sets available on Google Research for
your AI projects.
Scan QR code or visit:
http://research.google/tools/da tasets/

D omains of AI 27
 Maslow's Hierarchy of Needs: Maslow’s hierarchy of needs is a theory
in psychology proposed by Abraham Maslow in his 1943 paper "A Theory
of Human Motivation” in P sych o lo g ica l Review. It attempts to create a
classification system to reflect the universal needs of society as its base
and then proceeding to more acquired emotions.

Fig. 2.2 : Maslow’s Hierarchy of Needs

Cloud Storage is a model of data storage in which digital data is stored
in logical pools with the physical storage spanning multiple servers (can
be in multiple locations), and the physical environment is typically owned
and managed by an enterprise-hosting company.
Enterprise Resource Planning (ERP) is a business process management
software application that allows an organization to use a system of integrated
modules such as Finance, Sales, Inventory, etc., to manage the business and
automate many back office functions related to technology, services and
human resources. For example, SAP is an ERP. (Source: U n iv e r s it y o f L o u is ia n a )

A/B testing is a way to compare two versions of a single variable by

testing a subject's response to variant A against variant B and determining
which of the two variants is more effective.
Structured Query Language (SQL) is a standard Database (DB) language
used to create, maintain and retrieve the relational database. Some key
points to note about SQL are:
* SQL is case-insensitive. However, it is a recommended practice to use
keywords (like SELECT, UPDATE, CREATE, etc.) in capital letters and use
user-defined things (like table name, column name, etc.) in small letters.
* You can write comments in SQL using at the beginning of any line.
* SQL is the programming language for relational databases like MySQL,
Oracle, SQL Server, Postgre, etc. Other non-relational databases (also
called NoSQL databases) like MongoDB, DynamoDB, etc., do not use
SQL. These are new databases used in emerging technologies. Do read
about them.
* There is no ISO standard in SQL as such. You may encounter queries
that work in SQL Server but do not work in MySQL.

28 Essentials of A rtificial I ntelligence- I X

 • Big Data is an area within computing that looks at ways to analyze,
systematically extract information from, or otherwise handle data
sets that are too large or complex to be dealt with by traditional data-
processing application software. Data with many rows requires greater
statistical power while data with higher complexity (more attributes or
columns] may lead to a higher false discovery rate.

Computer Vision (CV)

Computer Vision (CV) has been around for over 50 years. Its development began
in the 1950s, around the same time when Artificial Intelligence gained prominence.
But like Al, it is only recently that CV has started gaining importance. Some CV
applications include self-driving cars, facial recognition-based tracking systems with
vision cameras and Amazon Go.

Interesting Read
Amazon Go is a new kind of store with no checkout
required. With the Just Walk Out Shopping experience,
simply use the Amazon Go app to enter the store,
take the products you want, and go! No lines, no
checkout. (Source: Amazon)

Scan QR code or visit:

youtube.com/watch?v=NrmMklMyrxc

The key underlying technology in these amazing use

cases is Computer Vision.

D omains of Al 29
 You can use the open source Computer Vision framework called
Open CV for your AI projects.

Scan QR code or visit:

https://opencv.org

Evolution of CV
In the summer of 1966, Seymour Papert and Marvin Minsky at MIT Artificial
Intelligence group started a project titled Summer Vision Project. The main aim
of this project was to build a system that could analyze a picture and identify all
objects in the image. It is interesting to note that even now CV is a complicated
area with thousands of researchers across the globe trying to perfect it but it was
started by two undergraduates who dared to solve it way back in the 1960s.
In the 1970s, David Marr, a neuroscientist at MIT, taking cue from brain studies including
cerebellum, hippocampus and cortex for human perception, set up the building blocks
of modern Computer Vision. He is known as the father of modern Computer Vision.
Majority of his research ideas are summarized in his book titled VISION.
Since 2012, Deep Learning, which is a subset of Artificial Intelligence, has gained
prominence. It focuses on Artificial Neural Networks (ANNs) which have been
covered in detail in Chapter 6. ANNs are designed on the lines of how neuron
networks work in our brain, so algorithms learn on their own from huge amounts
of structured or unstructured data.
Deep Learning can work hand in hand with CV creating powerful systems such as
searching images in Google, tagging of friends on social media, apps which can
create an aged version of your face, speech-to-text translation, intrusion detection
systems, etc.

30 E ssentials of A rtificial Intelligence - I X

Applications

Web: From traditional text searches, we have moved onto image search, face
recognition, object recognition, scene recognition, geolocalization from vision,
image tagging on social media, Google maps aerial imaging, YouTube (content
categorization)

VR/AR: Outside-in tracking, inside-out tracking (simultaneous localization and

mapping, HoloLens), object occlusion (dense depth estimation)

Smartphones: QR codes, computational photography (Android Lens Blur, iPhone

Portrait Mode), panorama construction (Google Photo Spheres), face detection,
expression detection (smile), Snapchat filters (face tracking), Google Lens, Night
Sight (Pixel)

Medical imaging: CAT/MRI reconstruction, assisted diagnosis, automatic pathology,

connectomics, Al-guided surgery

How Computer Vision Works

Computer Vision

Bowl, Orange,
Banana, Peach,
Mango, Apple

Input Sensing device Interpreting device

Bowl, Orange,
Banana, Peach,
Mango, Apple

Media: Visual effects for film, TV (reconstruction), virtual sports replay

(reconstruction), semantics-based auto edits (reconstruction, recognition)

Insurance: Claims automation, Damage analysis, Property inspection

(Source: T o w a rd s D ata S c ie n c e )

D omains of Al 31
Challenges

One would think that with so much advancement in technology, Computer Vision
is superior to human vision and can surpass it on all use cases. But this is not the
case. There are still lots of constraints in CV capabilities. When compared to human
vision in situations like simple face recognition under varying circumstances of
lighting, expression, additional objects/clothing on face, CV output is sometimes
not correct.

It also needs significant training data to be accurate compared to human vision.

Some of the other challenges with CV include:

Privacy and Ethics: The insurance industry is trying to customize motor

insurance premiums by analyzing user driving behaviour through surveillance
systems. On the flip side, Vision-powered surveillance systems can help identify
certain ethnic groups and minorities which is not desirable. Governments need
to take appropriate steps to tackle data privacy. European Union (EU) came out
with General Data Protection Rights (GDPR) for its citizens in 2018. Recently,
San Francisco has become the first city to ban the use of facial recognition by its
government.

Lack of Explainability: Modern artificial neural network-based algorithms are

still not clear in terms of how object classification is carried out. How self-driving
cars are designed algorithmically and their accuracy levels in object collision
avoidance, etc., are still not very clearly understood leading to ambiguity in laws
around it.

DeepFakes: Using deep learning techniques, it is possible now to create fake

images and videos of celebrities impersonating them and broadcast any message
or communication. Its use during election time can be very damaging to
democracy. In the US, the Pentagon, through the Defence Advanced Research
Projects Agency (DARPA), is working with research labs and institutions to tackle
DeepFakes.

Adversarial Attacks: Adversarial attacks are attacks where a hacker tries to

intentionally manipulate the inputs to machine learning models in order to cause
the model to make a mistake; they are like illusions for machines.

32 E ssentials of A rtificial Intelligence - I X

 Fig 2.4: Computer Vision model fails to recognize a person when a patch of paper is attached to him.
(Source: M e d iu m )

As per a recent report, the Computer Vision market was valued at USD 2.37 billion
in 2017 and is projected to reach USD 25.32 billion by 2023 at a Compound Annual
Growth Rate (CAGR) of 47.54%. (Source: M a rke tw a tch )

The world is undergoing a fundamental digital transformation that shows no signs

of slowing down. Average monthly data consumption of Jio alone is 10.8 GB.
According to a report, every minute:

• Users watch 4,146,600 YouTube videos

• Instagram users post 46,740 photos

• Snapchat users share 527,760 photos

All the above data sets give us a huge opportunity to apply Computer Vision to
derive patterns and analyze this data and develop useful applications out of it.

It must be pointed out that in spite of all these advancements in CV, Al and the
field of Computer Vision still need to tackle certain basic problems such as ethics,
bias, risk unawareness and lack of explainability.

D omains of Al 33
jLTzmmM
'Teachable Snake is an interactive web game powered by the beta version of Teachable
Machine 2 and React.js, inspired by Webcam Pacman project.
The idea is that instead of using physical buttons to control the game, the user can
draw a black arrow on a piece of white paper as controller and move the snake by
turning the paper in different directions in front of the webcam.
To recognize the paper controller pointing at different directions, the model is pre­
trained with 500 photos in each arrow direction (up, down, right and left) and stored in
cloud with the beta version of Teachable Machine 2. With the API endpoint published,
the model can be applied in the React.js app.
Scan QR code or visit:
https://experiments. with google, com/teachable-snake
V, J

Useful Resource
Full list of CV applications
Scan QR code or visit:
https://www.cs. ubc. ca/~!o we/vision. htm l

We are constantly surrounded by pictures, be it Google or your phone or

social media. People are taking selfies all the time and sharing on Instagram
or Snapchat.

The reason for this growth in images and videos is because of smartphones which
have cameras, and so taking a photo or video and sharing it has never been easier.
In fact, smartphones have overtaken cameras as the largest manufacturer of camera
lens. Earlier, when we used conventional cameras, each picture was worth money
and so people used their camera film judiciously. Now, if we don’t like a picture,
we can simply delete it or just add a filter to enhance it.

We all know that YouTube is essentially the second largest search engine with
hundreds of hours of videos uploaded every minute and billions of videos being
watched daily.

34 E ssentials of A rtificial Intelligence - I X

 What the computer sees

82% cat
__________ ________ ^ . 15% dog
image classification 2 % hat
1% mug

The internet consists of text, images and videos. We have been able to index and
search text for many years now and algorithmically also, it is fairly straightforward.
But for images in particular, we need to analyze what they contain and for that
we need Computer Vision. Over the years, meta tags and descriptions have proved
to be the best way to describe images but with CV, it is now possible to literally
“ see” and “ interpret” images and know what they contain.
This may sound complicated from a computer standpoint but it comes very naturally
to humans, even children.
Let us see the process of how we typically visualize and interpret vision:
• We can describe the contents of an image we see once.
• We can summarize an image or video that we have seen only once without any
prior training.
• We can recognize a face that we have seen before.
This is the bare minimum that we expect from Computer Vision as well. Doesn’t
sound very difficult, right?

At an abstract level, the goal of Computer Vision problems is to use the

observed image data to infer something about the world.
— C o m p u te r V ision: M odels, Learning, a n d Inference, 2 0 1 2

Computer Vision is sometimes confused and used interchangeably with Image

Processing.

D omains of Al 35
Image Processing is defined as
OUTPUT
the process of creating a new
INPUT
image from an already existing IMAGE DESCRIPTION
one, either simplifying or
enhancing the content in some IMAGE DIGITAL IMAGE
COMPUTER VISION
PROCESSING
way. It is basically a type of
digital signal processing and is
COMPUTER ARTIFICIAL
not concerned with interpreting DESCRIPTION
GRAPHICS INTELLIGENCE
the content of an image.

A given CV system may require Image Processing to be applied to its raw input,
e.g., pre-processing images.
Some instances of pre-processing of images include:
• Adjusting photometric properties of the image such as brightness or colour
• Cropping an image to centre an object in a photograph
• Reducing digital noise from an image such as extracting content from low light
levels

The goal of Computer Vision is to extract useful information from images.

This has proved a surprisingly challenging task; it has occupied thousands
of intelligent and creative minds over the last four decades, and despite this
we are still far from being able to build a general-purpose “seeing machine” .
— C o m p u te r V ision: M odels, Learning, a n d Inference, 2 0 1 2

Key elements of a CV application include but are not limited to:

• Object Classification: Category of object contained in the image such as human
or animal
• Object Identification: Type of object such as Cat or Dog
• Object Verification: Verify if the object is actually in the image
• Object Detection: Where exactly is the object in the image
• Object Landmark Detection: Key indicator points fo r the object in the image
• Object Segmentation: Pixels specific to the object in the image

• Object Recognition: Overall understanding of the object and where it is in the

image

36 E ssentials of A rtificial I ntelligence - I X

 Fig 2.5: Object Classification (Source: T o w a rd s D a ta S c ie n c e )

Deep Learning has fundamentally transformed CV capabilities. Let us see how.

Traditionally, in the Pre-Deep Learning era, CV was constrained with computational
capabilities. The typical process flow for simple tasks like facial recognition was:
• Create a database of all types of images of the subject in a particular format,
structure and with particular attributes (lighting, angle, etc.).
• Tag images with unique parameters such as eye distance, eye-to-nose length,
nose bridge, ear distance and many such attributes to uniquely define the image
and map to an individual person.
• Capture a new set of images to test the data set. You would gather the new
images and then go through the entire process of uniquely marking the images
on the above attributes. Then the application would run query on the database
and finally arrive at the conclusion— whether these images uniquely match any
in the existing database for recognition purposes.

As you can see, very little automation is involved here and the cost of recognition
would be very high as the database becomes larger, not to mention constraints
on capabilities due to lighting, angle, foreign objects in the images, etc.

Deep Learning provides a completely different novel approach to solving CV

problems. With Deep Learning, we no longer need to write complicated database
queries and manually code every single rule into their vision applications. Instead,
we can program “ features” — smaller applications that could detect specific
patterns in images. They then use a statistical learning algorithm such as linear
regression, logistic regression, decision trees or support vector machines (SVM)
to detect patterns and classify images and detect objects in them. Let us look at
some examples of CV in action.

D omains of Al 37
Facial Recognition Applications

Social m edia tagging, phone unlocking using face ID and attendance biom etric
system s using face— all of these use face recognition w hich is built on top of CV.
Law enforcem ent agencies are now increasingly using face recognition to identify
crim inals w ho change their facial features to go incognito. Passport agencies and
im m igration authorities use Face ID fo r visa purposes and border transit access.

Self-Driving Cars
It is not only technology com panies that are m aking use of CV, even the autom otive
industry is catching up. Self-driving cars about w hich we read in Chapter 1
fundam entally use CV to m ake sense of their surroundings. They have a 360-degree
cam era that captures the surroundings from different angles and then im ages are
pre-processed. This input is then fed into the CV console application which recognizes
the objects around, type of object, its distance, traffic signs, lane departure, etc.

Al algorithm s run on top of this CV console to provide trig ger to driving dynam ics
in order to control steering, speed and brakes. This helps drive carefully and
autonom ously w ithout hitting any objects or breaking any traffic rules.

Health Care
CV can analyze health records, lab reports,
X-ray im ages and prescription data/clinical
inform ation m uch faster than the doctors. It
can then run algorithm s against m illions of
em pirical records and do prognosis effectively.
An exam ple of CV applications in health care
is O rlando Health W innie Palm er Hospital for
W om en and Babies, w hich taps Com puter Vision through an artificial intelligence
tool that m easures blood loss during childbirth.
The key issue w as that post-partum (post giving birth) haem orrhaging w as one of
the biggest causes of m ortality during childbirth. The Al technology uses pictures
taken w ith an iPad device and analyzes im ages of surgical sponges and suction
canisters.

Since im plem enting Al at the hospital, w here 14 ,0 0 0 babies a year are delivered,
doctors have learned that they often overestim ate how m uch blood a w om an
loses during delivery. W ith Com puter Vision, they can understand the am ount m ore
accurately, allow ing them to treat w om en appropriately.

38 Essentials of A rtificial I ntelligence - I X

Mixed Reality (Augmented/Virtual Reality)

You m ust have seen Google Glass, Virtual Reality (VR) headsets and A ugm ented
Reality (AR) apps (look at the Fun Tim e section below ). These applications use CV
to em pow er a m ixed reality experience. They are used in education fo r learning,
in tourism industry fo r creating VR tours, and in oil m ining industries to capture
offsite areas w hich are difficult to reach or are unsafe fo r hum an beings.

• You can have a look at Google Glass here:

Scan QR code or visit:

^ tf)S:^ ww w ^ 0Utu^ e' com^watc^ ?v=4EvNxWhskf8

• See the magic of Hololens Mixed Reality:

Scan QR code or visit:

□ 9^1^ h ttp s://w w w .yo u tu b e . co m /w atch 7v=auJJrH gG 9M c

• See how VR can save lives by safety training:

Scan QR code or visit:

https://w w w .yo utu b e .com /w a tch ?v= p L qN b 8J7 \vlk& fe a tu re = e m b _ lo g o

A machine (NeuralCandy) that dispenses candy when the user displays the
requested image in front of the trained image model.
NeuralCandy combines image classifier and sugar
highs in one delicious Android Things project. ro Ht

The application asks for a random image to be

---------1
placed in front of the camera module and if it
•'li"
matches the request, then the motor of the candy -------------! » -

dispenser is activated to release the reward. ■ »* * • • • *•— ®i

NeuralCandy uses the TensorFlow Lite inference library for Android to locally classify
the captured image against the pre-trained ImageNet model. This model is good at
recognizing categories that it was trained with. You can use a smartphone to search on
Google for the requested target image and put it in front of the Pi camera. The Raspberry
Pi 3 model B will handle the image processing and the motor for the candy release.
' Scan QR code or visit:
https://experiments. with go ogle, com/neuralcandy

D omains of Al 39
 Teachable Machine is a web tool that makes it fast and easy to create machine
learning models for your projects, no coding required. Train a computer to recognize
your images, sounds and poses, then export your model for your sites, apps, and
more.

Scan QR code or visit:

https://experiments. withgoogle. com/teachable-machine

Natural Language Processing (NLP)

As we have seen earlier, unstructured data such as random com m ents, im ages,
videos and audios contain a lot of inform ation. Earlier, it w as com plicated and
expensive to m ake sense of it but w ith the advent of CV, it is easy to take care of
im ages and videos. But w hat about text and audio? Here's w here Natural Language
Processing (N LP) com es into play!

NLP is all about algorithm s to understand and interpret speech and text w hich is
natural language data. Be it any language, sentence or paragraph, NLP can scrape
through it and m ake sense. Traditionally, hum ans interact with each other in natural
language but w hen it com es to interacting w ith m achines, we use program m ing
languages and softw are codes.

W hat a fun w ould it be if we could talk to m achines in the sam e w ay as we talk

to other people! NLP m akes it possible.

To v isu a liz e , a ty p ica l interactio n between h u m a n s and m a c h in e s u sin g

N atural Lang uag e P ro c e s sin g is as follow s:
1 . A h u m an ta lk s or w rites to the m ach in e (bot).
2 . The m ach in e ca p tu re s the au d io or text.
3 . Audio-to-text conversion or any language tran slatio n for text-to-text takes
place.
4 . P ro c e s sin g of the text’s data to u n d erstan d and extract intent (sig n a l
c la s s ific a tio n ) follow s.
5 . Now d a ta -to -a u d io conversion for audio or any language tra n sla tio n for
text-to-text takes place.
6. The m ach in e rep lies to the hum an by p la yin g the aud io file or renderin g
the text in the interface.

40 Essentials of A rtificial I ntelligence - I X

Som e of the w ell-know n applications of NLP we use in our everyday lives are:
• Translation tools such as Google Translate, M icrosoft Translato r

• Docum ent Processors such as M icrosoft W ord and G ram m arly that em ploy NLP
to check fo r gram m atical, sem antic errors and plagiarism to check accuracy of
texts

• Standard Interactive Voice Response (IVR) applications used in call centres to

handle support queries
• Personal assistant applications such as Google A ssistant, Siri and Alexa
• A new article sum m arization app called Inshorts w hich converts a full-length
new s article to a 6o-w ord new s brief along with a headline and a card im age.
(Source: E T )

• There is an interesting merging of CV and NLP in an application called WordLens.

Take a look at it:

Scan QR code or visit:

https://w w w .youtube. com /w atch 7v=h20fQdYrHRs

You can use the Apache Open source framework Apache OpenNLP for
your AI projects.

Scan QR code or visit:

https://opennlp.apache.org

Every new technology com es with its pros and cons. NLP is no different. There are
constraints to NLP as w ell. As sentences becom e longer, and as we use slangs and
incorrect gram m ar or m ultiple languages in the sam e content, the intent becom es
hard to extract and classify. This leads to errors in interpretation. A hum orous
exam ple of such a fallacy is:

In 1950s, English to Russian translation w as m uch sought after due to Cold W ar.
Here is a biblical sentence that required translation: “The spirit is w illing, but the
flesh is w e ak.”

D omains of AI 41
 SYNTACTIC ANALYSIS— GRAMMAR
sentence

noun phrase verb phrase

noun phrase noun phrase

det noun conj det noun verb det noun

the boy and the girl ate their tea

Here is the result w hen the sentence w as translated into Russian and back into
English: “ The d rin k is g o o d , b u t th e m ea t is r o tte n .”
(Source: N e w Y o r k T im e s )

NLP at its heart uses two underlying logics: Syntax-based analysis and Semantic analysis.

Let us see how it is used.

1. Syntax

Syntax m eans how the w ords are organized in a sentence in the proper order and
using correct gram m ar so that they m ake sense.
NLP needs to understand the structure of the sentence and breaks it dow n into
m ultiple objects to feed into the sem antic analysis engine.
Let us take an exam ple text:
Ashish w o rk s a t A g n itio . Ashish is an A rtific ia l In te llige n ce P ra ctitio n e r. He s tu d ie d a t
th e Indian In s titu te o f Technology and w o rk e d a t Nestle.
Som e of the com m on syntax-based analyses are:

• Sentence Segm entation: This m eans to sim ply break a paragraph into sentences.
Each sentence represents a separate idea or concept.
From the above exam ple, we get:
“ Ashish w o rk s a t A g n itio .”
“ Ashish is an A rtific ia l In te llig e n ce P ra c titio n e r.”
“ He s tu d ie d a t th e Indian In s titu te o f Technology and w o rk e d a t N estle.”
• W ord To ken izatio n : This m eans breaking the sentence into w ords.
From the above exam ple, we get:

“ A shish ” “ w o rk s ” “ a t” “ A g n itio ”
Just split the w ords w hich have spaces betw een them . Easy!

42 Essentials of A rtificial I ntelligence - I X

• P re d ictin g Parts of Speech fo r Each Token: This m eans looking at each w ord
individually and understanding the w ords around it to interpret if it is a proper
noun or verb or adjective. W e can also m odel this part to create an input-output
schem a here.

From the above exam ple, we get:

Input M odel O utput

W ord token: “A sh ish ” Speech Part A shish is a proper

Surrounding w ords: classifier m odel noun
“w o rks”

“a t” “A g n itio ”

• Text Lem m atization: This m eans to figure out the basic lem m a of a w ord.
For exam ple,
I have one baby
I have tw o babies
These are tw o separate w ords but their lem m a is sam e and one is basically the
plural of another. •

• Id e n tifyin g Stop W ords: There are lot of w ords like “T h e ”, “an d ”, “ is” in the
English language w hich are filler w ords and NLP engines need to have a m odel
to filte r them out.

Google has developed a tool to empower computer systems to read and understand human language in
order to process it in intelligent ways. Google decided to share the findings with the community by releasing
SyntaxNet, an open-source neural network framework for TensorFlow that provides a foundation for Natural
Language Understanding (NLU) systems.

Scan QR code or visit:

https://ai. googleblog. co m /2 0 16/05/announcing-syntaxnet-w orlds-m ost.htm l

p Scan QR code or visit:

https://github.com /spoddutur/syntaxnet
0 f «

D omains of Al 43
2. Semantics
Sem antics m eans trying to understand the core intent of the sentence in order to
reply appropriately. It involves applying com puter algorithm s to understand the
m eaning and interpretation of w ords and how sentences are structured.

Here are som e techniques in sem antic analysis:

• Nam ed Entity R eco gnitio n (N ER): It involves understanding w hich parts of a
text can be identified into preset or predefined set of groups w hich could be
nam es of people or places of interest, etc.
A typical NER system can tag parts like this:
* Nam es of People * Nam es of C o untries/G eograp hic Locations
* Currency * Dates and Tim es
* Historic Events

• Play with SpaCy, the leading open-source library for advanced NLP, and Prodigy, an annotation tool for
radically efficient machine teaching.

Scan QR code or visit:

https://explosion.ai/dem os/displacy-ent •

• D ependency Parsing: This m eans to understand the dependency and correlation

betw een various w ord tokens across sentences.
For exam ple, in the first line “A shish w orks at A g n itio ”, “A sh ish ” w ord token is
a proper noun. Sim ilarly, in the second line “A shish is an Artificial Intelligence
P ractitioner”, “A sh ish ” w ord token is again a proper noun. In the third line,
“ He studied at the Indian Institute of Technology and w orked at N estle”, it is
to be interpreted that “ He” m eans “A sh ish ”.

• W ord Sense D isam b ig u atio n : It involves giving m eaning to a w ord based on

the context. A w ord m ight have m ultiple m eanings and the sense it has m ay
change depending on the context.
For exam ple,
I am on a “ c a ll” .
You have to “ c a ll” her.
Both sentences use the w ord token “ca ll” but the m eaning is very different in
both the cases.

44 Essentials of A rtificial I ntelligence - I X

• Coreference Resolution: By now, we understand various parts of speech, how
they are interrelated, w hich are the nouns and NER. One issue we need to
look at is use of pronouns w hich are repeated in a text. NLP engine needs to
understand w hich pronoun refers to w hich noun from the context and state
of conversation flow .
For exa m p le , Ashish is a reader. A shish likes to go w ith Salil fo r w ork. He also
m eets other people at w ork.
Here, it m ight be am biguous to a sim ple algorithm w hether “ He” is fo r A shish
or Salil. Coreference Resolution is the technique to help solve this and identify
w hich pronoun is referring to w hich noun.
• Natural Language Generation: It involves using databases to derive sem antic
intentions and convert them into hum an language.

■Purpose: To understand how three Al domains are interrelated to each other.

You know that Al has three distinct sub-fields— Natural Language Processing,
Computer Vision and Data for AL They are very different from each other but only
together do they constitute the whole concept of Artificial Intelligence.
Take three different colour strands and work
them into a braid. See how long your braid
can become within 30 seconds! Ready? Go!!
To understand Al, we draw an analogy with the three strands in a braid. One is the
Data strand, the second is the Natural Language Processing strand and the third strand
is Computer Vision. Together, they constitute the concept of Artificial Intelligence.

•Objective Type Questions :

A. Fill in the blanks:

1 ..................................... means how the words are organized in a sentence in correct
order and grammar so that they make sense.
2. ... ............ means to understand the dependency and correlation between
various word tokens across sentences.
3 . Al is the next step from ....... . ............in search.
4 . Amazon Go u se s.................................. to capture product details and their prices in a
cashierless store environment.
5 . News summarization apps use ............. to make a summary of large news
articles into small pockets of information.

D omains of Al 45
 6. Breaking sentence into words in NLP is called .....................................
7 ....................................means trying to understand the core intent of the sentence in
order to reply appropriately.
8 ....................................is the first step in order to do data visualization.

B. State whether the following statements are True or False.

1 . In earlier days, use of unstructured data required heavy computation power.
2 . Al systems can work better with little data as well.
3 . Big Data, NLP and CV can't be integrated together.
4 . Computer Vision has surpassed human vision as of today.
5 . NLP can be used for assessments of subjective questions without human intervention.
6. Using Computer Vision, face recognition on phones is possible.
7 . Now, data in the form of media files such as images and videos can also be processed
for Al systems.
8. loT and Al can be integrated together.

C. Multiple Choice Questions (MCQs):

1 . What technology do tools like Alexa and Siri use to understand voice commands?
(a) Augmented Reality (b) Natural Language Processing (NLP)
(c) Neural Networks (d) Blockchain
2 . What is a challenge with Computer Vision technology?
(a) Perceptron (b) DeepFakes (c) Skewness (d) Normality
3 . What underlying technology does a self-driving car use for object recognition?
(a) Computer Vision (b) Neural Networks
(c) Augmented Reality (d) Natural Language Processing (NLP)
4 . When did Machine Computer Vision surpass human vision accuracy?
(a) 2010 (b) 2015 (c) 2019 (d) 2000
5 . What is the process of converting unstructured raw data with variances into structured
data?
(a) Data Cleansing (b) Data Shining
(c) Lemmetization (d) Natural Language Processing (NLP)
6. What technique in semantic analysis involves determining the parts of a text that can
be identified and categorized into preset groups?
(a) Named Entity Recognition (NER) (b) Supervised Learning
(c) Lemmetization (d) Annotation
7 . What type of data set constitutes images and videos?
(a) Structured Data (b) Level Data
(c) Unstructured Data (d) Responsive Data
8. Which is not an example of Computer Vision?
(a) Siri (b) Photos tagging in social media
(c) Face ID in phones (d) Mixed Reality

46 Essentials of A rtificial I ntelligence - I X

 9 . Adjusting photometric properties of the image, such as brightness or colour, is
known as:
(a) Data Cleansing (b) Image Recognition
(c) Preprocessing (d) Natural Language Processing (NLP)
10 . "I have a car but my brother has two cars." Here, car and cars are similar words and
NLP captures them into one common entity. What is this entity called?
(a) Lemma (b) Token (c) Schema (d) NER

Subjective Type Questions •

1 . What do you think are the possible challenges with using Computer Vision in
E-Governance?
2 . Describe how you could use Natural Language Processing in subjective assessments
in school.
3 . Describe how you will build a chatbot that uses NLP and can talk to mental health
patients and make them feel better.
4 . Do you think Al and Big Data are related and dependent on each other? How?
5 . Imagine you are running a call centre. You are facing a major problem—your staff is
not in their seats most of the time with the result a lot of calls remain unanswered.
How will you use CV, NLP and Data together to solve this problem?
6. A lot of data is generated at your school every day. Classify this data into various
categories and think what kind of data visualizations can you run on it?
7 . Will it be possible to use NLP in future for teaching? What would be its limitations
and possible areas it can be used for?
• 8. Think of examples where Data, CV and NLP are integrated in common-life applications.

^ ^ ^ u i c k Activity
i i i i i i i
Le t t e r to Fu t u r e S elf

Put on your reflective cap and write a letter to your future self. What do you want to tell yourself or remind
yourself about?
Imagine the world in 2030 and write a letter to your future self. Be sure to mention things that you think
your future self would probably be doing and experiencing in daily life.
Place........................................
Date........................................
Dear Future Self,
It feels a little weird to write to you but I want to share how life is now so that you can relate to it in the
future.
Hope this letter brings back old memories...
Sincerely

D omains of Al 47
 Spen d Tim e w it h G o o g le St o ry Speaker

Build an interactive story with Google Story Speaker and discuss with your friends.
• Install Story Speaker extension for Story Speaker:
S&HKB
jg K jg Scan QR code or visit:
https://chrom e. google, com /w ebstore/detail/storyspeaker/ohfibfhhfbhknfdkipjdopbnegkbkjpj

• Introduction to Story Speaker:

Scan QR code or visit:

h ttps://w w w .youtube, com /w atch ?v=wsrzvYYvhH8& feature=youtu. be

• Read more about Story Speaker:

Scan QR code or visit:

https://docs.google.eom/document/d/1hFrBtsBbF2LoZ1FFpXEH7L6fWH1lj24W1-itXnKSXK8/edlt

• Basic Template of Story Speaker:

Scan QR code or visit:

https://experiments. with google, com/story-speaker

The Story Speaker extension can be accessed with docs.google.com. After signing in on Google and opening
Google Docs, go to Add-Ons and choose Story Speaker. A dialog will open. Click on Basic Template and work
on the same after understanding how it works.
(Source: C B S E h a n d b o o k )

Using your Floor Plan (Chapter 1, Activity 4) as a base for your story, use the Story Speaker extension in
Google Docs and write the story. Ensure that you give/write specific instructions in the story to enable the
storyteller to respond.
Now it is your turn to create a story using the Google Extension of Story Speaker for Google Docs.
START HERE
Link for the basic template of Story Speaker:

Scan QR code or visit:

https://experiments. with google, com/story-speaker

48 Essentials of A rtificial I ntelligence- I X

 Applications of Al
in S olving
Real-World Issues
Learning Objectives:
1^ 1 To appreciate complexity of real-world issues
[ e j To identify leverage points which are appropriate for achieving Sustainable Development
Goals (SDGs) using Al solutions

B ackground and C o n text

“ ...Climate change is a massive problem across nearly every sector and measure of
human development. To address it at the speed and scale that current conditions require,
we will need to take a more data-driven approach— one that harnesses the full power
of Artificial Intelligence and other advanced technologies to accelerate discovery and
innovation at a truly planetary scale.”
— Lucas Joppa, Chief Environmental Scientist, Microsoft

From creating the perfect crop to early stage prognosis of diseases to inclusive
quality education fo r all, Al can actually help in solving global internally developed
M illennium Sustainable D evelopm ent Goals (SDGs).

Al FOR SOCIAL GOOD

Artificial Intelligence is frequently portrayed in apocalyptic term s as the technology
that will take over our jobs or even our lives or at least m ake hum an beings a
slave to technology. But w hat if Al could also becom e a valuable tool in w orldw ide
efforts to achieve goals of our society?

In 2017, after the passage of Hurricane Harvey, m any streets in dow ntow n Houston
w ere flooded and im passable— but others w eren't. An Al application that used
satellite-based im agery w ith object detection and identification algorithm enabled
the rescue team to identify safer escape routes fo r those trapped by the rising
w ater levels.

A pplications of Al in S olving R eal-W orld I ssues 49

A i-p o w ered object detection could bring relief to visually im paired people, m ost
of w hom live in developing countries. The app w orks through a sm artphone and
uses Al to recognize know n people and their attributes as w ell as specific objects
such as currency bills.

These are just tw o small exam ples of Al being used fo r social good— and we have
scratched the surface on the possibilities of A l. Although Artificial Intelligence is
not a panacea, it could help in solving som e of the w orld's m ost challenging social
problem s such as curing cancer and climate issues. Al doesn't always require scientific
breakthroughs; it can add to existing efforts in both developed and developing
econom ies which are experiencing significant societal challenges or crises and often
live beyond the reach of traditional solutions that w orked in the past.

Fig. 3.1 : Al as agent of the future world society

Several Al capabilities are applicable to a w ide range of societal challenges; however,

they are m ore pertinent in the dom ains of health and hunger, education, security
and justice, and equality and inclusion w here typically a large population w ould
be affected. In the field of health, fo r exam ple, as per M cKinsey Global Institute,
people afflicted w ith diabetes could be helped by an A l-enab led w earable device
that can detect potential early signs of diabetes through heart rate sensor data.
Al could also analyze m obile phone im ages to distinguish betw een a benign skin
lesion and m elanom a, potentially helping m illions of rural dw ellers w ho do not
have access to derm atologists.

50 E ssentials of A rtificial Intelligence- I X

In education, adaptive learning technology, w hich tailors content fo r students
based on their abilities, could benefit around 1.5 billion students w orldw ide.

Fig. 3.2 : Al and Social Good (Source: P e w R e s e a r c h C e n tre )

A l AND SDGs
In 2 0 15 ,19 3 countries of the w orld agreed to the United Nations (UN) 2030 Agenda
fo r Sustainable D evelopm ent to set up goals fo r the betterm ent of the w orld.
These countries identified 17 global goals to end poverty, protect the planet and
ensure peace and prosperity for people, now and in the future. These 17 Sustainable
D evelopm ent Goals (SD Gs) are defined fo r joint global cooperation betw een the
developed and developing countries.

1 NO QUALITY GENDER
1 POVERTY 2 ZERO
HUNGER
HUNC
GOOD HEALTH
AND WELL-BEING 4 EDUCATION EQUALITY
CLEAN WATER
AND SANITATION
• • •

I V* I III I f I 9
•i

AFFORDABLE AND DECENT WORK AND ^ REDUCED RESPONSIBLE

CLEAN ENERGY
>• /
ECONOMIC GROWTH J INEQUALITIES 2 CONSUMPTION
AND PRODUCTION

' 1' t t i ® OO
U LIFE LIFE 1 c PEACE, JUSTICE PARTNERSHIPS
14 BELL
BELOW WATER ON LAND 10 AND STRONG FOR THE GOALS
INSTITUTIONS

K
Fig. 3.3 : UN’s 17 Sustainable Development Goals (SDGs) (Source: im t.o rg )

A pplications of Al in S olving R eal-W orld Issues 51

 Sustainable Development Goals
1 . No Poverty: Bring advances in Al to the most vulnerable to ensure algorithmic
equity
2 . Zero Hunger: Use Al to improve agriculture for better food security

3 . Good Health and Well-being: Standardize a framework for the performance

benchmarking of 'Al for Health' algorithms to address health issues
4 . Quality Education: Education can enable students to thrive in a world increasingly
augmented by Al-powered technologies
5 . Gender Equality: Actionable and thoughtful ways to identify and address gender
inequalities
6. Clean Water and Sanitation: Tackle the challenges such as lack of expertise,
climate change, resource optimization, and consumer trust
7 . Affordable and Clean Energy: Improve photovoltaic energy capture

8. Decent Work and Economic Growth: Industrial Al to accelerate economic growth

9 . Industry Innovation and Infrastructure: Bridge the digital divide and increase
. .
connectivity gains
1 0 . Reduced Inequality: Align all stakeholders in civil society and improve
transparency
1 1 . Sustainable Cities and Communities: More resilient, innovative systems in
smart cities
1 2 . Responsible Consumption and Production: Promote the use of Al and other
frontier technologies such as loTs, Big Data analytics, and 5 G
1 3 . Climate Action: Take advantage of frontier technologies in combating climate
change and achieving a circular economy
1 4 . Life Below Water: Advanced deep-sea technologies for autonomous, fast, high-
resolution ocean exploration
1 5 . Life on Land: Al-powered drones to help fight ocean plastic and poachers

1 6 . Peace, Justice and Strong Institutions: Digital transformation to ensure

connectivity
1 7 . Partnerships with Goals: Bring diverse stakeholders from around the world
together and develop skills for the digital economy and society
(Source: U N )

52 E ssentials of A rtificial Intelligence- I X

A ch ievin g SDGs across the w orld is not just a m oral prerogative but an econom ic
imperative. These 17 goals are integrated in such a m anner that action on one will
im pact others. W hile the w orld is m aking good progress in som e areas, we are
falling behind in delivering the SDGs overall. We need all stakeholders— businesses,
governm ents, academ ia, international forum s, non-profit organizations and others—
to com e tog ether and scale their efforts to deliver the SDGs using every tool at
th eir disposal, including A rtificial Intelligence.

In Decem ber 2017, the 203oVision published its first report: U niting to Deliver
Technology fo r the Global Goals. In this report, role of digital technology— big
data, robotics, internet of things, Al and other technologies— w as identified in
achieving SDGs. In this chapter, we will discuss how Al can be an effective tool in
accom plishing alm ost all of these SDGs.

SDG 1: NO POVERTY

R ecognizing the causes of poverty is key to tackling the

problem using Al technology. From natural disasters,
w ar and conflict, unavailability and unaffordability of
food, lack of life skills, education— Al can help identify
the reg io n ’s m ost needy. The possible solutions involve
im proving farm ing lands and agriculture, increasing
education and helping the inhabitants learn new skills
to support com m unities in distribution in poorer and
w ar-torn areas, or w here natural disasters have w reaked
havoc. Al can be used in all these cases.

To identify the m ost needy individuals, groups and regions, im ages taken by
satellites on a continuous stream can help identify global activities that reflect
poor and rich regions. For exam ple, geographic areas w ith a high light density at
night-tim e tend to be typically w ealthier than those w hich are dark w ith little or
no access to electricity at night. A team of researchers from Stanford U niversity
used satellite im ages of areas in daylight and night-tim e im ages to identify the
m ost poverty-stricken regions of Africa. They w ere able to predict poverty w ith an
81-99% accuracy as com pared to the household survey data. Al can help identify
areas w hich are badly in need of help.

A pplications of Al in S olving R eal-W orld Issues 53

Improving Agriculture
Al can help farm ers analyze a variety of things in real tim e such as w eather
conditions, tem perature, w ater usage or soil conditions collected from their farm s
to m ake inform ed decisions. Al technology helps in detecting diseases in plants,
pests, and poor plant nutrition on the farm s; in preventing over-application of
herbicides and excessive toxins that find their w ay in our food and in creating
seasonal forecasting m odels to im prove agricultural accuracy and predict upcom ing
w eather patterns m onths in advance.

Fig. 3.4 : Al enabling smart agriculture

Improving Education Levels

Poverty and lack of education go hand in hand. Al can help im prove education
levels (both access and quality) in poor areas. Intelligent chatbots as teachers for
students w ithout access to other form s of schooling, are not very far off. This A l-
teach e r could guide students through a syllabus and assess the learning levels of
individual students using real-tim e analytics. It could eradicate the m oney barrier
in accessing quality education. Tailored learning through Al addresses different
learning preferences of individuals instead of the current m ethods of just one form
of learning and creates a level playing field.

SDG 2: ZERO HUNGER ZERO

The w orld will soon w itness a hunger crisis. A ccording HUNGER
to the United Nations, by 2050, we will need to increase
the w o rld ’s food production by 70% to feed the w o rld ’s
population. Even today, one in eight people in the w orld
w
do esn ’t get enough food to eat.

54 E ssentials of A rtificial Intelligence- I X

Al system s could be deployed for sm arter and m ore efficient food production.
Traditionally, food production is increased by expanding cultivable land. But this
cannot be done any m ore as nearly half of the potentially productive farm land has
already been used fo r agriculture. C onsidering the risk of fragile ecosystem and
the im pact of global w arm ing, it is necessary that w e go beyond the traditional
w ays to solve the w orld hunger problem .

Fig. 3.5 : Precision Farming (Source: E n te r p r is e i o T In s ig h t s )

Growing the Perfect Crop

Al tech n ology can help grow seeds w ith right genetic m ake-up that generate the
highest yield, have m ost nutrition and are largely disease-resistant. Al w ill process
all the grow th, genetic and environm ental data. It will review all the param eters
and varieties to identify the patterns and insights faster than hum ans. Besides,
m ore efficient plants will be chosen fo r production.

Maximizing Current Output

At present, a large quantity of the w o rld ’s food ends up in trash. Al can separate
food m ore specifically as “g o o d ” and “ bad.” It can evaluate food to determ ine the
best w ay to use it. For exam ple, a tom ato m ight not be good enough to m ake it
to the salad plate, but it w ould be perfect to be used in m aking ketchup or soup.

Precise Application of Herbicides

Rather than “sp ray-an d -p ray” approach to herbicide application, Al can be used to
distinguish betw een a w eed and a sprout of a plant based on learning from m ore
than a m illion im ages of young plants. W hen it identifies a w eed, it can spray on
it directly. This can cut losses by up to 90%.

A pplications of Al in S olving R eal-W orld Issues 55

Early Warning of a Crisis

Al can help identify the areas w hich are at an increased risk of food shortages due
to lack of w ater, rising food prices and crop failure. Such a system has already been
deployed in Colom bia to w arn the farm ers of a potential drought and suggest them
to skip the planting season. The farm ers w ho heeded the advice and skipped the
planting season, saved on planting costs w hen the drought occurred.

SDG 3: GOOD HEALTH AND WELL-BEING

W hen it com es to health care and w ell-being, Al has
the potential to provide data-driven clinical decision
support (CDS) to physicians and hospital staff by
identifying patterns, using algorithm s and data to
provide autom ated insights into health care providers.
Second, an Al system can be generated that predicts the
outcom es of hospital visits. This prevents readm issions
and reduces the duration of patients’ stay in hospitals.
Al can im prove health care by fostering preventive
m edicine and new drug discovery. It can also help by identifying specific treatm ent
fo r patients.

Integrating Al into health care ecosystem has m any benefits including autom ating
tasks and analyzing big patient data sets to deliver better health care faster and

56 E ssentials of A rtificial Intelligence- I X

cheaper. A ccording to Business In sid e r , 30% of health care costs are associated
w ith adm inistrative tasks. Al can autom ate cognitive tasks like pre-authorizing
insurance based on rule engine, follow ing up on outstanding bills and m ain tain in g
health records to ease the w orkload of health care professionals and ultim ately
saving them tim e and m oney.

Al can pull tog ether patient insights leading to a predictive analysis. It can also
help in obtaining patient insights to discover key areas of patient care that require
im provem ent.

W earable health care technology using Al helps serve patients better. Algorithm that
uses A l, like FitBits and sm artw atches, can analyze health data and the underlying
patterns to alert the users and their health care providers on potential health risks
in future. A ssessing o n e ’s own health through technology eases the w orkload of
professionals and prevents unnecessary hospital visits or rem issions.

SDG 4: QUALITY EDUCATION

One of the goals of the United N ations is to ensure
quality education for all and prom ote lifelong learning.
Easy access to internet has already m ade education
accessible and more uniform around the globe. However,
Al can bring drastic changes in the existing education
system in the follow ing w ays:

Personalized Learning

In traditional education system , the curriculum is designed to suit as m any learners

as possible by targeting 80% of the norm al distribution curve. How ever, students
struggle to attain their full potential w hen they are in the top 10% or bottom 10%.
They have difficulties follow ing along.

How ever, w ith the introduction of Al, teachers can offer personalized assistance
to each student based on their learning needs and the kind of learners they are
(text, audio, video, action-oriented). Al can custom ize hom ew ork assignm ents as
well as final exam s, ensuring that students get the best possible personalized
assistance.

A pplications of Al in S olving R eal-W orld Issues 57

Intelligent, Supportive Learning Environments
Research indicates that instant feedback is one of the keys to successful tutoring.
But in a conventional classroom setting, teachers typically do not have the tim e
fo r one-on-one tu toring and feedback. Instead, w ork is set to be review ed later,
leading to a lag betw een effort and feedback. This dilutes the learning experience
and m akes it harder fo r the teachers to tailo r content to their stud en ts’ needs.
Through A l-p ow ered apps, students are targeted, custom ized and get real-tim e
responses from their teachers.

Simplifying Administrative Tasks

Al can autom ate adm inistrative tasks for teachers and academ ic institutions.
Teachers spend a lot of tim e grading exam s and assessing hom ew ork. Al can be
used to autom ate the grading tasks w here m ultiple tests are involved. This w ay,
teachers can spend m ore tim e on their students.

Smart Content
Al can produce sm art content of superior quality w hich includes virtual content
like vid eoconferencing and video lectures. Textbooks are already taking a new
turn. Al system s can create custom ized textbooks. As a result, textbooks are being
digitized and new learning interfaces are being created to help students of all
academ ic grades and ages.

Al in Education
Problems Al Solutions
Standardized curriculum does not cater to Personalized learning
----------- ►
individual needs
Limited one-to-one tutor time available for Personal virtual tutors
university students
Grading and assessment is time-consuming, Al can assess open-ended questions— and in
with an over-reliance on multiple choice real time
Large class sizes in K-12 schools means Virtual classroom assistants
----------- ►
children’s questions often go unanswered
Personalized communication is almost Chatbots can answer administrative questions
impossible due to scale from parents, staff and students on the fly
Selecting the best students from a large Al can shortlist candidates based on multiple
application pool “ 1* data points
Increasing drop-out rates at universities ------------ ► Al sentiment analysis
The need to effectively combat plagiarism and Natural language processing can identify
ensuring authorship patterns and source facts
Fig. 3.7: Al in Education (Source: E d u T e c h A sia )

58 E ssentials of A rtificial Intelligence- I X

SDG 5: GENDER EQUALITY

Can A rtificial Intelligence help bridge the traditional

gap betw een the genders? M ost studies show that
gender discrim ination will perpetuate the bias in Al and
m ore parity m ust be ensured in the society before Al is
fully adopted. But this has a flip side too. A l, to som e
extent, can help in bridging gender gap too.

Progress has ...but challenges remain

been made... Across 149 countries assessed, there are large
disparities in political empowerment...
Gender parity in education is
W o m en represent
almost complete
17 heads of state
T h e r e is h i g h e n r o l m e n t i n s e c o n d a r y e d u c a t i o n g l o b a l l y . . .
18% of ministers
24% of parliam entarians
65% of girls 34% of managers
66% of boys ...as well as economic empowerment

...but low attendance in college/university. in just 60% of countries studied, women have
as much access to financial services as men.

3 9 % of women in 42% of countries, women have as

e
much access to land ownership as men.
34 % of men women spend double the time men do on
housework and other upaid activities in the
and 20% of women are illiterate in 44 countries. 29 countries for which data are available.

And new gender gaps are

emerging in the jobs of the future
Women are sorely underrepresented in Al and other careers that Even in the countries with the largest Al talent pools, there is a
require science, technology, engineering and math skills. significant gender gap among Al professionals.

Al workforce distribution
Women make up 22 per cent of Al workers, but the gender gap varies Share of Remaining
professionals with gender gap
by industry. Al skills by gender (female/male)
In d u s try (g e n d e r g a p )

S o f t w a r e a n d IT S e r v ic e s ( . 2 3 ) f f f j f l l 7 .4 % 3 2 .5 %
•► 70%
E d u c a t io n ( . 3 3 ) W H W W M W W W W t W t 4 .6 % 1 3 .9 %

F in a n c e ( .2 4 ) fttW 1 .4 % 5 .7 %

M a n u fa c tu r in g (.1 8 ) f l i j f j f 10% 54%

•►72%

o
C o r p o r a t e s e r v ic e s ( .3 0 ) m 1 .0 % 3 .5 %

H a r d w a r e a n d n e t w o r k i n g ( .2 3 ) tut 0 .8 % 3 .6 % F e m a le M a le
84%
H e a lth c a r e ( .3 4 ) Iff 0.9% 2 6 . %
Germany
I______ I______ I______ I-------- 1-------- 1
10 20 30 40 50
■ M a le ■ F e m a le
S h a r e o f A l t a l e n t p o o l (% )

Fig. 3.8 : Gender Disparity (Source: im f )

A pplications of Al in S olving R eal-W orld I ssues 59

Addressing Wage Disparity

A rtificial Intelligence can be used to address

the gender pay gap w hich exists in m ost places.
Often, com panies try to quickly solve their
gender pay gap problem s by offering high fresher
salaries and other benefits to attract m ore fem ale
applicants. But if inequalities exist in levelling
and later prom otions and pay structures, this
w o n ’t result in a long-term solution. Al can
rem ove discrepancies in a m ore sustainable way
that d o e sn ’t disadvantage either gender. Using m achine learning algorithm s,
they are able to m erge and analyze the data together providing expertise in three
key areas:

• Identifying existing pay discrepancies based on gender, ethnicity, disability or

any other factor.

• Insights into the causes behind these gaps, based on data and academ ic
expertise from the backend.

• O pportunities to take data-driven decisions to narrow these gaps.

W ith the use of A l, com panies can address the problem of salary-based inequalities
and, for exam ple, identify discrepancies which could arise from less obvious factors
like leave travel allow ances, m aternity benefits, housing and car allow ances, etc.
U sing this inform ation, com panies can zero in on w hich functions or at w hich
levels real gaps exist and use data-driven approach to take decisions to create
real actionable change.

Detecting Online Sexual Harassment

Al program m es can be designed to m onitor internal com m unications such as

com pany docum ents, em ails and chats, and flag inappropriate content and bring
it to the notice of em ployers. How ever, som e Al experts say that such technology
has som e challenges such as protection of data privacy and that Al can only look
fo r certain triggers but cannot pick up on broader cultural or unique interpersonal
dynam ics.

60 E ssentials of A rtificial Intelligence- I X

SDG 6: CLEAN WATER AND SANITATION
The challenge of ensuring availability of safe drinking
w ater for all is at the heart of SDG 6. Sanitation and CLEANWATER
hygiene are key related issues since the lack of these may ANDSANITATION
lead to m any diseases, especially am ong young children.

Al can be used fo r detecting any dangerous bacteria

and harm ful particles in w ater. You w ould be shocked
to know that about 2 billion people use a drinking w ater 9
source contam inated w ith faeces. A prototype device
w as m ade that uses pattern recognition and m achine learning to m onitor and
inspect w ater quality. The application also m arks the contam inated sites on a map.
For regions in the w orld w here access to clean drinking w ater is a challenge, sim ple
test system s like this could dram atically help in preventing diseases and saving
thousands of lives.

A rtificial Intelligence can be used fo r m ore effective w ater treatm ent processes as
well as fo r detecting problem s and assuring an early allocation of efforts. Al can
offer accurate and tim ely inform ation about the kinds of pollutants in w ater and
m ake recom m endations fo r their treatm ent.

SDG 7: AFFORDABLE AND CLEAN ENERGY

It is well know n that with the rise in per capita incom es,
global energy dem ands are grow ing every year and
conventional fossil fuels w on’t be able to fulfil our energy
needs in the future. Em issions from fossil fuels hit an all-
tim e high in 2018 due to increased energy consum ption
and pow er requirem ents.

There have also been calls for making emission norms more
stringent in the automotive industry and factories globally.

On the other hand, renew able energy is rapidly being accepted as a reliable
alternative to fossil fuels. Look at the grow th of electric vehicles around us.
These resources are m uch safer and cleaner than conventional sources like petrol,
diesel, etc. W ith the advancem ents in technology, renew able energy sector has
m ade significant progress in the last decade especially in term s of its reliability,
production, efficiency and cost.

A pplications of Al in S olving R eal-W orld Issues 61

How ever, there are still a few challenges to this sector that can be addressed with
the help of em erging technologies. A rtificial Intelligence and M achine Learning
technologies can analyze the past, optim ize the present and predict the future.

SUSTAINABLE
DEVELOPMENT
G O ALS
ENSURE ACCESS TO
AFFORDABLE, RELIABLE,
SUSTAINABLE AND MODERN
ENERGY FOR ALL □
G LO B A LLY
-----------

12.6%
STILL LACK ACCESS TO
MODERN ELECTRICITY
ENERGY ACCOUNTS
FO R RO UG HLY

2/3
OF GLOBAL GREENHOUSE
GAS EMISSIONS
Fig. 3.9 : Access to Clean and Affordable Energy (Source: U n ite d N a tio n s, In d ia )

Smart, Centralized Control Centres

The energy grid can be interconnected w ith loT devices and sensors to collect a
large am ount of data and interpret it to m ake sense.

W hen com bined w ith the pow er of A l, this data can give new insights to the grid
operators for better control over operations by offering flexibility to the energy
suppliers to sm artly adjust supply w ith dem and. Advanced control system s such
as industrial furnaces, boilers or large air conditioning units, can be installed with
equipm ent w hich can autom atically sw itch off w hen pow er supply is low or certain
scenarios are achieved. Intelligent storage units can also be optim ized based on the
flow of supply. Al m achines and sensors can m ake w eather and load predictions.
This will im prove overall integration and efficiency of renew able energy.

62 E ssentials of A rtificial Intelligence-IX

Improved Integration of Microgrids

Using A l, we can do integration of m icrogrids and m anage distributed energy.

The A l-p ow ered control system can play a vital role in resolving quality and
congestion issues.

Improved Safety and Reliability

Al can m anage irregularity in the energy sector and also offer im proved safety,
efficiency and reliability. It can trace energy consum ption patterns, identify energy
leakage and m onitor health of the devices. For exam ple, in production units, the A l-
powered predictive analysis can collect data to m onitor w ear and tear, m onitor overall
health of the equipm ent and alert the operator when the m aintenance is needed.

Expanding the Market

A l-pow ered system s can provide insights into energy consum ption by analyzing data
on energy collection. This w ould help suppliers optim ize existing services, launch
new service m odels and help retail suppliers to target new consum er m arkets.

Smart Grid with Intelligent Storage

Sm art Grid w ith Intelligent Energy Storage (IES) can analyze a vast am ount of data
captured from several loT sensors and m ake tim ely decisions on energy allocation,
helping m icrogrids to efficiently m anage local energy needs while continuing pow er
exchange w ith the m ain grid.

SDG 8: DECENT WORK AND ECONOMIC GROWTH

Skills for Jobs of the Future
There is a large gap betw een the jobs and skills available.
Young people are generally engaged in low -skilled jobs DECENTWORKAND
w ith few er opportunities fo r perm anent sustainable ECONOMICGROWTH
em ploym ent as education system rem ains out of step
w ith the needs of rapidly changing w orkplaces. Low-
skilled jobs are losing to autom ation. So, it will be critical
that young people are equipped w ith the necessary
f i t
skills to secure quality w ork. Al can bridge this gap
betw een the desired and delivered skills, through
tailored learning.

A pplications of Al in S olving R eal-W orld Issues 63

Financial Access

Globally, tw o billion people lack access to form al financial services such as banking,
transactions, investm ents, etc. W hen people from underprivileged backgrounds
gain access to basic financial services (e.g., savings, investm ent), they take the
first step tow ards greater security. Al can analyze data to identify people in need
of loan or those w ho are unable to access it. Al can also identify good and bad
debtors based on previous transaction data. How ever, those looking to start a
new business, lack of credit history can be a barrier.

Insufficient Labour Demand

By 2030, job landscape will look vastly different from today as the econom y will
w itness profound evolution or disruption through digital technology. M ost of the
children entering prim ary school now will w ork in jobs that do not exist today. Al
technology could create job opportunities fo r highly skilled w orkers. It can help
bridge the gap betw een labour dem and and supply m ism atch.

SDG 9: INDUSTRY, INNOVATION AND INFRASTRUCTURE

Structural Defects

Al is being used to identify problem s w ith structures or

INDUSTRY, INNOVATION
equipm ent before they cause critical system s’ failure.
O rganizations can use Al to detect cracks or faults in
design or concrete or broader issues in the structure,
9 ANDINFRASTRUCTURE

highlighting the severity of the problem . By identifying

and segm enting an issue, engineers can figure out the
exact shape, size and scale of the issue, and associate
it w ith other pieces of critical inform ation to form ulate
a solution before the problem goes out of hand.

Landscape Designing

As we have learnt in the previous tw o chapters, autonom ous vehicles and equipm ent
can use CV to identify objects in front of them and take appropriate action according
to the object type. Using A l, it can learn and understand to identify the typical
behaviour of other users of cars such as direction, distance and speed of approach
and take appropriate steps to avoid it.

64 E ssentials of A rtificial Intelligence- I X

In everyday life as well, we see sm artphones deploying Al and ML for object
identification and learning user behaviour. Such system s can be used to design
and create, for exam ple, upgrading or creating roads w ithout trees on either side
of the road, etc.

Waste Management
Al can also be used for w aste m anagem ent, a problem staring into our faces.
Intelligent trash can, equipped w ith Al program s, RFID tags, loT sensors, is a
revolutionary concept in the w aste m anagem ent sector. This can be used to
m easure w aste levels of the garbage throw n inside them and send data to the
main disposal system for processing. The system identifies the type of the garbage,
the quantity of each type of garbage and the respective w aste disposal m ethod.
G arbage dum ps have started deploying w aste-sorting robots w hich can sort tons
of garbage tirelessly in a day.

Smart Recycling

Al-pow ered sm art bins can think for them selves while sorting and sending garbage.
W hen garbage is put into the bin, it uses sensors to study trash and decides by itself
w hat needs to be done w ith the garbage. It sends the garbage to the appropriate
disposal system , be it a dum ping ground or a recycling factory. This way, there
can be a significant decrease in the w aste generated globally.

Al in garbage disposal and smart recycling

—
8________________________________________________________/

Al reads the tag to determine the type of

garbage and sends it to the main server
for analysis

The main server compares the various

garbage categories and picks an
appropriate disposal or recycling method

Safe disposal or clean recycling of the

garbage is carried out
V____________________________________________ /
Fig. 3.10: Al in G arbage D isp o s a l (Source: A iie rin )

A pplications of Al in S olving R eal-W orld Issues 65

S D G 1 0: R E D U C E D IN E Q U A L I T I E S

G row ing inequality is a challenge fo r developed and

developing countries alike. The incom e disparity has
g ro w n w id e r o ver th e years. A cro ss O ECD , a verage
incom e of the richest 10% of the population is roughly
nine tim es that of the poorest 10%. This is up seven
tim es as com pared to w hat it w as 25 years ago. Not
just w ithin a country, inequality betw een countries is
also grow ing. From 1960 to 2016, the gap betw een the
average incom es of people in the richest and the poorest
countries grew by 135%.

Top earners continue to pull away from everyone else

Cumulative per cent change in real hourly wages since 2000

+25% Top 5%

_10 % ------------------------------------------------------------------------
2000 2018
Fig. 3.11: Growing Income Inequalities (Source: E c o n o m ic P o lic y in stitu te )

Tailored Technology
Tailored technology products and services to serve the needs of low -incom e and
vulnerable groups can be a significant m ove to address inequalities. The goal is to
m ake them m ore affordable and accessible fo r rem ote populations w hich can be
achieved w ith Al such as by creating m obile tariffs and service bundles focused
on increasing affordability.

Social Media and Crowdsourcing Platforms

Such platform s can encourage participation and inclusive engagem ent. Social
netw orks can support positive activism and cam paigns aim ed at solving critical
social challenges. Com m unities can report problem s such as dam aged roads and
graffiti, prom ote participative design of solutions to social needs such as schools
and infrastructure.

66 E ssentials of A rtificial Intelligence- I X

Helping Refugees
Al program s can provide psychological support to people w ho have experienced
shock and traum a w hether physically or em otionally. Chatbots can help o rganizatio n s
connect and respond to displaced people. Al can help refugees prepare before
departing by providing inform ation about national borders and transportation and
advice on security. Al technology can also help them em bed in the new society.
A ccess and connectivity allow them navigate local services such as housing and
em ploym ent, access opportunities, ensure safety, etc.

Automating Justice and Equal Employment

Com panies or governm ents can use m achine learning to predict good candidates
to hire or the likelihood of repeat offences to decide on sentencing, w hich, in turn,
w ould reduce individual biases.

Creating Accessibility
New education opp ortunities are popping up every day, m any of w hich are free
or low -cost and provide equal training fo r this constantly developing industry.
These low -cost options rem ove entry barriers created by private (or even public)
institutions allow ing low -incom e or underprivileged people access these benefits.

Helping Visually Impaired to Navigate their Environment

By providing navigational assistance via sm artphone, object recognition and reading
text in any language, converting handw ritten or printed text into digital text so
that it can be spoken aloud using text to speech converters, Al can m ake life easier
fo r the visually im paired.

SDG 11: SUSTAINABLE CITIES AND COMMUNITIES

Sm art Cities (Indian Governm ent also started the
concept of Sm art Cities a few years back) w ould
give m any benefits to society, including enhanced
opportunities fo r education, im proved skills and,
hence, job prospects, better access to health care
and cleaner drinking w ater.

There w ill also be a lot of societal and environm ental

challenges w hen cities start to becom e sm arter.
Inefficient planning and non-inclusive m anagem ent

A pplications of Al in S olving R eal-W orld Issues 67

practices lead to certain unsustainable settlem ents that do not enable people to
advance em otionally, socially and econom ically. Sm art and innovative technologies,
including Al, are revolutionizing the w ay cities can address the challenges associated
w ith sm art hyper grow th.

Maximum Utilization of Existing Resources

Al technologies can help cities utilize and allocate existing assets m ore effectively
and efficiently. It can also help im prove how inform ation is m anaged and shared
across system s. Satellite data is transform ing the w ay cities are m anaged. From
identifying pow er outages to sm art routing for traffic m anagem ent and allocation
of resources during disasters, Al can truly help in optim izing a city and also in
connecting and integrating m ultiple cities.

Planning and Zoning

Al can help in m anaging urban governance, fo r exam ple, w hen green spaces are
converted into built-up structures. Al system can look for satellite im ages to quickly
reveal how the city developm ent aligns with original planning and zoning layouts
or w hich com m unities are m ost prone to flooding or natural disasters. Al can also
help detect and m ap different types of land cover, land use and type across space
and tim e, and generate im portant insights, analytics and data visualizations.

Planning for Scalability and Sustainability

Al system s can detect feasibility of upscaling activities to other locations or cities,
consider the sustainability of the project and how the hardw are/equipm ent m ay
be used during and after it has been com pleted.

SDG 12: RESPONSIBLE CONSUMPTION AND PRODUCTION

The key goal here is to change the w ay we produce
and consum e goods and services. It is im portant to
acknow ledge that production and consum ption
goods and services drive econom ic grow th and im prove
of 12 RESPONSIBLE
CONSUMPTIONAND
PRODUCTION
quality of life. But this should be done w ith few er
natural resources as consum ption of resources im pacts
environm ent across the life cycle. The focus is to do o o
m ore with less: use m inim um resources and create least
w astage.

68 E ssentials of A rtificial Intelligence- I X

Business Model and Product Innovation

By studying consum er behaviour through A l, business m odels can be designed

to optim ize consum ption and reduce w astage. Sm art refrigerators, equipped
w ith sensors and vision cam eras accessible through sm artphone apps, allow
users to check their contents w hile shopping. This helps reduce w astage of
food w hich, based on consum ption levels, can be recorded autom atically.
Sim ilarly, sm art buildings and offices can help users optim ize their energy
consum ption.

To look up green buildings,

Scan QR code or visit:
h ttp s://w w w .y o u tu b e .c o m /w a tc h ? v= E S IH w iV 8 l 6 l<

Precision Recycling

A l-enab led robots can ensure that all

products, including nanoscale m aterials,
are recovered and reused or recycled. Metal
detectors, precision w eight m easurem ent,
3D scanning and printing and spectral
analysis enable m ore sophisticated and
precise w aste sorting.

Materials Tracking Systems

Use of RFID (Radio Frequency Identification Device) and other loT sensors
can provide sophisticated tracking of m aterials to optim ize planning and m ovem ent
of the m aterial in the supply chain. Al can go through all the data and predict
various stages of disposal and lim it illegal disposal of hazardous m aterials and
w aste.

3D Printing

It can im prove the efficiency of m aterials by sim plifying the processes and avoiding
w aste that com es w ith m ass production. It prom ises longer life span of products
by m aking it possible to produce replacem ent parts on dem and.

A pplications of Al in S olving R eal-W orld I ssues 69

SDG 13: CLIMATE ACTION
Climate change impacts all countries of the world.
However, the poor and vulnerable population is
disproportionately affected. Artificial Intelligence can
help us understand the current climate situation, draw
patterns and predict future weather events to create
new products and services to minimize human impact
on environment.

Climate Study

By analyzing data from sensors, gauges and monitors, machines can spot patterns
quickly, accurately and automatically. The more accurate our current climate stats
are, the better climate models can be designed. We can also identify our biggest
vulnerabilities and risk zones.

Developing Better Solutions

Artificial Intelligence can help climate researchers and innovators to test theories
and solutions as well as “ what-if” scenarios that involve environment-friendly
measures. By using data provided by machine learning algorithms, organizations
have been able to cut the amount of energy used or reduce their carbon
footprint.

Green Initiatives

The easier we make green initiatives for people while making them feel responsible,
the higher is the adoption rate. Artificial Intelligence and machine learning algorithms
can help curate products and services that are easy to adopt. There are several Al
devices such as smart thermostats and smart irrigation systems that help conserve
energy resources. If we all do our individual contribution, our planet will be greener
very soon.

Better Weather Event Predictions

Early warning signs of catastrophic weather events can reduce damage to human
lives and property. Historical data across geographies is available for analysis. Al
has made significant progress in using machine-learning algorithms to assess the
duration and severity of these catastrophic events.

70 E ssentials of A rtificial I ntelligence - I X

SDG 14: LIFE BELOW WATER
As you know, the earth is covered almost 70% by
water and our oceans are home to a wide variety of
flora and fauna. With the advancement in satellite
imagery and machine learning, we are able to
analyze this massive amount of data to help protect
underwater life.

Monitoring and Regulation of Illegal Fishing

Intelligent systems using motion sensors, GPS and cameras can increase
accountability in the fishing industry by detecting anomalies and make it more
efficient for regulators and boat owners to review footage.

Traceability and Identification of Species

Tracking and monitoring of species will be easier and efficient with digital tools
like facial or species-recognition software. RFID and barcode scanners in supply
chains can help prevent sale of endangered or vulnerable species.

Smart Aquaculture

With the growing demand for fish protein, reliance on aquaculture to become
smarter will increase. Expanding wild fish stocks is no solution. Aquaculture
presents an opportunity to track fish and seafood from “ farm to plate" with the
help of intelligent tracking systems and loT sensors. This can inform the market
about limiting or increasing fish stocks based on demand and help in reducing
seafood fraud which is widespread. Sensors can also detect and optimize water
quality parameters and streamline feeding protocols.

Environmental Monitoring and Smart Conservation

loT sensors can collect data on temperature, salinity and pH which can be developed
into usable information using machine learning algorithms. This will help in planning
and balancing the needs of fishermen, transport and ocean conservation. Tools
can be used to monitor fish, track conservation progress and make well-informed
decisions like limiting the capture of unwanted marine lives.

A pplications of Al in S olving R eal -W orld Issues 71

Tracking Fishing Vessels
Al uses satellite-based monitoring system to track all fishing vessels in real-time
which can help protect fisheries around the world. They can collect satellite imagery
and analyze movements of boats with a specially designed form of machine learning
algorithms to determine if they are fishing vessels. This can be posted on to their
open source platform so that researchers, government, law enforcement agencies
and the general public can keep an eye on any fishing boats that venture into
protected waters by mistake.

Ocean Cleaning
Al can also be used in ocean cleaning. Satellite imaging can be used to detect
polluted patches; sensors can be used to identify the type of pollution— oil spills,
plastic, etc. Using smart systems, this waste can be suitably treated or recycled.

SDG 15: LIFE ON LAND

The key challenges to protecting life on the earth include
addressing species and habitat extinctions, which are
driven by climate change, pollution, unsustainable
exploitation, invasive species and habitat conversion
due to agriculture. Climate change is a major threat to
ecosystems around the world, clean water availability,
habitat and biodiversity which, in turn, increases our
vulnerability to climate change.

Preventing Wildfires

Climate change threatens to increase the number and size of wildfires. Al algorithms
can help in early detection, help firefighters combat blazes and aid in recovery.
Al can trace the pattern of climatic or anthropogenic events, eventually leading
to wildfire.

Checking Deforestation

Al can be used to track deforestation in real time in remote communities. Sometimes,

seeing the change in trees is easy but working out why it is happening requires
complex pattern recognition. For example, forests could be cut down because
they are a planted crop such as palm oil or lumber. Forests could be destroyed by
weather events, illegally cut for lumber, firewood or most commonly cleared to
make way fo r agriculture. So, Al and ML methods need to be used to figure out

72 E ssentials of A rtificial Intelligence - I X

which of these causes is at play and what is their individual impact. Al is yielding
optimal consumption and production levels with vertical green farms, eliminating
waste and vastly improving yields and resource efficiency.

Spotting and Stopping Poachers

Foot patrolling and drone-based surveillance have not been effective in preventing
poaching. These efforts are labour-intensive and costly. Al-based methods have got
some initial success in combating poachers. This method uses image classification and
object detection to find animals and poachers on infrared video captured by a drone.
Artificial Intelligence can be deployed to catch wildlife poachers.
Six steps from offline training of Al model to online detection

3. User interface
A neural network is trained on Drones are flown over wildlife Video and still images are
70 videos containing animals sanctuaries, capturing thermal transmitted via radio waves
and poachers, which have been infrared images. to a computer.
labelled. The model is tested
with other videos.

4. Preprocessing 5. Detection 6. Output

Infrared images may need to be The video is processed in The neural network outputs
converted into “white-hot” format, batches and sent to the cloud for annotations that are overlaid
where warm objects are lighter analysis. Each image is treated on top of the original image.
against a dark background. as an input into the neural This enables identification of
network. the poachers’ whereabouts

F i g . 3 . 1 2 : A l t o s p o t a n d S t o p p o a c h e r s (Source: McKinsey & Co)

A pplications of Al in S olving R eal -W orld Issues 73

SDG 16: PEACE, JUSTICE AND STRONG INSTITUTIONS
This goal focuses on promoting peace, justice, and
accountable and inclusive institutions. Some countries 1 £ PEACE, JUSTICE

K
still face prolonged armed conflicts and a large population I D ANDSTRONG
is living under security threat. Various forms of violence INSTITUTIONS
against children, human trafficking, bribery, etc., persist ^ s'/.
worldwide. Steps need to be taken to make transparent
and traceable supply chains, improve effectiveness and
transparency in governance, create safer communities,
promote freedom of expression and opposition, reduce
cybercrimes and economic espionage, etc. Digital technology will play a key role
in handling these challenges.

Blockchain
It is an emerging tool that promises great trust and traceability. Blockchain can
solve problems such as blood diamonds, forced labour and ivory poaching. It can
be used to track the origin of products ranging from coffee beans to cotton.
Blockchain may also eliminate fraud and misappropriation of resources in areas
where governance is weak and corruption rampant.

Intelligent Cybersecurity
Al can be used to address the increasing risk of cyber attacks. Products with
adaptive algorithms that can study large data sets to better anticipate and predict
threats learn from previous attacks to give real-time response to attacks.

Automating Processes
Al deployment in processes such as tax collection, welfare distribution and
designation of land rights, etc., where information is recorded electronically on a
distributed ledger technology (such as Blockchain) and with access controls in place
so that it can be accessed only by those who need it, will lower the risk of corruption.

E-Governance
App-based platforms and chatbots can help citizens connect and engage with
governments and provide transparency and better accountability of government.
Access to information, contacting local politicians, tracing people, companies and
assets across the globe are a few things that can be achieved with Al.

74 E ssentials of A rtificial Intelligence - I X

Detecting Suspicious Behaviours
The real-time analysis of transactions, use of anomaly detection and risk-scoring
algorithms alongside predictive modelling can help in the identification of illicit
financial flows and payments.

SDG 17: PARTNERSHIPS FOR THE GOALS

Goal 17 aims at renewed partnership across governments,
civil society, private sector, the United Nations system
and other actors. The aim is to mobilize resources to
address all the goals. It also aims at supporting developing
countries fo r equitable progress by increasing capacity,
collaboration and knowledge-sharing, enhancing use of
enabling technologies and improving data monitoring
and accountability.

Digital Collaboration Platforms

Collaborative software using machine learning can make decision-making within
and between institutions more efficient. Through information exchange, education
and training platforms can improve capacity-building. Al can enable partnerships
to plan, implement, manage and improve initiatives to meet the individual needs
of partners around the globe.

Measurement and Tracking of Goals

loT and Al will enable measurement and tracking of progress towards the goals
which is crucial for improved decision-making. Al is also expected to improve
analytical abilities and guide us in decision-making in a smart and sustainable
manner keeping in mind data security, personal and emotional needs of the citizens
and safety of the planet.

Applications of Al at Present

Although a lot remains to be done, Al has already been applied for non-commercial
purposes by various companies and organizations.
Planet Labs, an earth-imaging Silicon Valley startup, partnered with Paul Allen
(co-founder of Microsoft), a few philanthropists and leading research scientists
to create a map of shallow water coral reefs by applying object detection to
satellite imagery in correlation to geospatial data. The map is used to monitor
reef ecosystems that are under threat.

A pplications of Al in S olving R eal -W orld Issues 75

At Thorn, an international anti-human trafficking non-profit organization, a
combination of Computer Vision, NLP and Big Data, is being used to identify
victims of sexual exploitation on the internet. It works with a group of technology
companies, including Google, Microsoft, and Facebook, and has identified a total
of 5,791 victims since 2016.

Researchers at the MIT Media Lab have applied reinforcement learning-based

algorithms in clinical trials with patients diagnosed with glioblastoma (a form of
brain cancer) to successfully reduce toxic chemotherapy and radiotherapy doses.
This is very encouraging because reducing chemotherapy doses helps improve
quality of life of cancer patients and cuts cost of their treatment. As research
continues to improve reinforcement learning, practical applications of solutions
will extend beyond clinical trials to customization of patient treatment.

A South Korean city, Songdo, deploys RFID tags to classify garbage categories.
A pressurized garbage disposal system can read this tag from a distance and an
appropriate method to dispose of total waste generated is decided dynamically.

Smartfin is another great example of an innovative surfboard fin with ocean health
sensors and a community initiative intended to galvanize communities to become
custodians of coastal ecosystems. It integrates temperature, motion and GPS
sensors into performance surfboard fins so that surfers can collect data simply
by doing what they love. Sounds amazing, no?

Microsoft's Seeing Al is accessible to users in 70 countries around the world and is

free of charge. Similar solutions include OrCam MyEye camera, which is mounted
on standard glasses and can convert visual information input into spoken output.
The device is portable and operates without the need for a smartphone.

There are fintech startups such as ZestFinance and Lenddo that capture data
by device, browser and social media trail to generate a predictive model of
creditworthiness through various parameters. M-Shwari banking, which leverages the
M-Pesa mobile money system in Kenya, incorporates history of telecommunications
in its assessment of credit risk. One in five adults in Kenya is currently an active user
of the service, and M-Shwari is regarded as one of the most successful solutions
for financial inclusion. Besides, its SMS and internet-based interface, predictive
algorithms leverage several Al capabilities to analyze social and telecom data and
assess their creditworthiness. The information is then processed in minutes and
produces a credit score, which determines the size of the loan to be sanctioned.

76 E ssentials of A rtificial Intelligence - I X

Researchers at the University of Southern California’s Centre for Artificial Intelligence
in Society built the SPOT system which automates the process of detecting
poachers through infrared video feeds and increases the reliability of surveillance.

Challenges Before AI

Scaling up AI use for social good will require overcoming some significant bottlenecks
like data accessibility as, in many cases, such data capturing may not be commercially
viable or data may be unstructured. Hence, data is not readily accessible to social or
non-governmental organizations. Bureaucratic inertia keeps useful data locked up
with government agencies. In other cases, the required data has never been collected.
AI capabilities need to be continuously improved but availability of talent with
high-level expertise in AI is in short supply. In other cases, implementation and
program management becomes a challenge as well. Non-profit organizations may
lack skilled resources such as data scientists and NLP engines needed to address
the AI tools and techniques. Sometimes, these tools may be prone to misuse by
authorities and others having access to them.
Therefore, the principles for their use need to be established. Human bias may
be embedded in AI models or data sets that could amplify existing inequalities.
Data privacy needs to be protected to prevent sensitive personal information from
being made public. AI applications need to be safe for human use and must comply
with law and social norms. Social acceptance, transparency and quality checks still
remain areas of concern.

^Objective Type Questions :

A. Fill in the blanks:

1. AI can overcome.............................. barrier in accessing education.
2.......................... a n d ........................ are examples of wearable health care technology.
3. Key challenges to protecting life on Earth include..................... and ..........................
4. AI can trace the pattern of both........ ................... and............................. events
leading to wildfires.
5. AI deployment in processes such a s ...............................a n d ................................ can
lower the risk of bribery and corruption.
6. Human biases are embedded in AI amplifying existing......................................

A pplications of AI in S olving R eal -W orld Issues 77

 7. Smart Grid w ith ....................................can analyze a vast amount of data captured
from several loT sensors and make timely decisions on energy allocation.
8. Intelligent systems using ............................, ........................... and ...........................
can increase accountability in the fishing industry by detecting anomalies and make
it more efficient for regulators and boat owners to review footage.

B. State whether the following statements are True or False.

1. Al-based systems can help detect crime using CCTV footage proactively.
2. Al can be used to track deforestation in real time in remote communities.
3. Average income of the richest 10% is roughly five times that of the poorest 10%.
4. Al can help detect bullying and sexual harassment online and on social media.
5. Al can enable satellite-based monitoring system to track all fishing vessels in real time
which can help protect fisheries around the world.
6. True Personalized Learning can be enabled through Al.
7. Al-assisted Clinical Decision Support (CDS) can replace doctors.
8. According to the United Nations, by 2050, we will need to increase the world's food
production by 70% to feed the world's population.

C. Multiple Choice Questions (MCQs):

1. Read the given statements:
I: Al can solve challenges which are beyond traditional solutions.
II: Drawback with Al is that it always requires scientific breakthroughs. Identify the
correct statements.
(a) I (b) II (c) Both I and II (d) None of these
2. Which of the following is not a feasible option any more to improve agricultural
productivity?
(a) Precision farming (b) Genetically modified seed varieties
(c) Expanding area under agriculture (d) Using climate-resilient crop varieties
3. Al can change current education models. However, which of the following cannot be
substituted by Al?
(a) Grading and assessment (b) Content creating and content delivery
(c) Accessibility and affordability (d) Emotional assistance
4. Which of the following is an understandable ground for wage disparity among
teachers in a school?
(a) Gender (b) Ethnicity
(c) Physical disability (d) Academic merit
5. Blood diamond, forced labour, ivory poaching and tracing origin of products. Which
of the following technologies is best suited to solve this issue?
(a) Satellite imaging (b) Robotics
(c) Blockchain (d) Deep learning

78 E ssentials of A rtificial I ntelligence - I X

 6. Which of the following are the challenges before Al? (More than one option possible)
(a) Talent availability (b) Scalability
(c) Data availability (d) Social acceptance
7. Which of the following is not an example of using Al in managing energy needs of
the world?
(a) Integration of microgrids
(b) Tracking of energy consumption patterns
(c) Predictive analysis of energy equipment health
(d) Logging of energy data
8. Al and loT can be used for facial recognition to understand species and trace them
below water. Which SDG do they impact?
(a) SDG 10 (b) SDG 14 (c) SDG 5 (d) SDG 9

: Subjective Type Questions :

1. Productivity is a big challenge in agriculture. How can Al solve this?

2. How can Al address hunger issues at present? Mention any two ways.
3. What are the challenges to marine life?
4. What do you understand by Intelligent trash cans?
5. How is Al helpful in infrastructure development?
6. You have read how Al can be helpful in meeting SDGs. Can you think of instances
where Al is an inhibitor to any of these goals?
7. Discuss any three applications of Al in health care.
8. Transparency is a challenge in present-day governance in many countries. How can
Al address this problem?
9. List any two disadvantages of Al.
10. What are the barriers to Al realizing its full potential? List any four.

r '
_________ * L _________ __ __________________
*

Quick Activity r~

Al FOR IMPROVING COOKING OPERATIONS

Read how Akshaya Patra, the largest NGO-run school meal programme, with support from Accenture Labs
improved cooking operations with the Al system.

If you could design an Al system, which social problem will you target? How will this system function? Are
there any negative impacts of this technology?

A pplications of Al in S olving R eal -W orld Issues 79

pgjjnSD
Please note these facts:
The United Nations estimates that we will need a 70% rise in food production to
feed the world by 2050.
According to the UK’s Institution of Mechanical Engineers, as much as half the
world’s food (or two billion tonnes) ends up in trash. Approximately 20% -40% of
crop losses are caused by pathogens, animals and weeds.
In the US, for example, farmers use 310 million pounds of herbicide on corn, soy
and cotton fields.
Africa’s famine crisis is the gravest emergency since the Second World War, according
to the United Nations, with six million people at risk of starvation in Somalia and
14 million more in South Sudan, Nigeria and Yemen. In addition, one in four persons
in sub-Saharan Africa is malnourished.
With this background, can you think of an Al-powered solution to these problems?
v__________________________________________________________________________ ,
BB35BSB.. ....... ”.. ...
GO-GOALS BOARD GAME
Link to Download Go-Goals Game Material:
https://go-goals.org/downloadable-material/
Let your imagination flow and make your own tokens for the Go-Goals Game! Think
of your favourite superhero or your favourite cartoon character and draw it on your
token. Cut it out, shape it, and win the Game!
You can also make your own dice using the DIY kit! Just cut it out, fold it and paste it!
Great! You are all set to begin!
Go-Goals Board Game
This is what your game board looks like. We are going to play in groups of 5. Have
you played ‘Snakes and Ladders’ before? This game is very similar to ‘Snakes and
Ladders’. Players advance the number of spaces by rolling a single dice. If a player
lands at the bottom of a ladder, they can immediately climb to the top. If a player
lands at the top of the slide, they immediately move to the bottom of the slide. If a
player lands on SDG goal field (1-17), they will draw a card corresponding to the
goal number. Another player will read the card question. A correct answer from the
card drawer will allow the player to roll the dice again. The first player to reach ‘2030’
is the winner. Let’s Play!”

80 E ssentials of A rtificial Intelligence - I X

A pplications of Al in S olving R eal -W orld Issues
 The purpose of the
Go GOALS game is to help you
understand the Sustainable
Development Goals, how they
impact your life and what you can
do every day to help achieve the
IT Goals by 2030. Let's play together
to make the world a better place!
www.go-goals.org
'W e have read about Future Skills in Chapter 1. We need to now research how jobs
will evolve to use those future work skills. This research can be used to help us
decide what we want to learn to be ready for these jobs. As there are numerous
jobs that will be created due to Al, we will conduct this research in teams! Each
team will be given a theme, some guiding resources and laptop to research on
Al-related jobs. The themes include Health, Security, Education, Entertainment,
Service, Transport. You will search for current and emerging trends in employment
and fill in the Research Template given in the handbook. The company can even be
a start-up, big organization or comm unity project.
Make even number of teams and divide, give one theme to two teams. For example,
if you have 10 teams, divide any 5 themes giving same theme to two teams. This
will help in the activities of next modules.

Put your Research Skills in action:

Here is a set of themes— you will be getting one theme out of these to work on in a
group of 4. Research on the given theme and find out various organizations incorporating
Al in your theme sector. Also, look at the skill-sets required for such jobs today. With
this, try to analyze what skill-sets would be required 10 years down the line?

Table Format:
What are the names of the organizations in our country Write briefly what they
working around this theme? do.

What kind of skill-sets (look at both soft skills and technical skills) do they look for
in their new hires?
Would the skill-sets required today still be the same after 10 years?
Soft Skills (Today) Technical Skills (Today)

Soft Skills (10 years later) Technical Skills (10 years later)

82 E ssentials of A rtificial I ntelligence - IX

 What are the ethical concerns revolving around the theme?
(Keywords: Al ethics, Al bias, Al access, Al privacy)

Topic Examples

Once we have completed this activity, we have to prepare a Job Advertisement for
2031— 10 years from now!
V_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _________________________________________

Job Advertisement for 2031!

Now, on the basis of the research you have made, make a “Future Job Ad” for 2031.
Your Job Ad should include information about the company that is hiring and what
kind of skills they are looking for in their employees. Share the reasons why you
chose this job or jobs. Be creative about the Job Ad you are making. Remember,
it is something from the future. What jobs will be relevant 10 years from now?

A pplications of Al in S olving R eal -W orld Issues 83

 Let’s begin.
Be as creative and sound futuristic in your tone as possible!
What to do:
• Search for current and emerging trends in employment to make a Future Job
Advertisement.
• The job description is for a job which will occur ten years from now, i.e., the
current date.
• To help you, the job advertisement must include the following information.
■ Required Skills: Technical and Soft Skills
■ Company Background: Is the company a big organization?
■ Analysis: Share the reason why you chose this job or jobs.
List the kinds of futuristic job opportunities that would be available for you.
Write the skills you will require to do these jobs.
Create a Job Advertisement for the Future below:

84 E ssentials of A rtificial Intelligence - I X

 o
# * :/> ♦

' © - E t h ic s in Al
f S

Learning Objectives:
Ip | Understand and appreciate ethical concerns around Al.
Ip | Critically think about the cost and benefits of Al technology.

B ackground and C ontext

“ In this era of profound digital transformation, it’s important to remember that business, as
well as government, has a role to play in creating shared prosperity— not just prosperity.
After all, the same technologies that can be used to concentrate wealth and power can
also be used to distribute it more widely and empower more people.”
— Erik Brynjolfsson, Director of the MIT Initiative on the Digital Economy

Fig. 4.1: Understanding Ethics (Source: Machine Learning, XKCD)

E thics in Al 85
'W a tc h the video: Al for Good
jgSfl&p Scan QR code or visit:
httpsV/www.youtube. com/watch ?v=vgUWKXVv09Q

What have you understood from the video?

What are your learnings from it?
V._________________________________________________________________________________ /

Rallnnn Debate:
Balloon nr
Divide yourselves into groups of four. Two groups will be
given a theme related to advantages and disadvantages of
various Al applications in different industries you researched
about. One team will speak in favour of Al applications while
the other will speak against it. The debate will go theme by
theme and each member of both the teams will get a minute
to speak. The chance to speak will alternate between the favouring
and rebuttal team. Finally, only one team will remain in the balloon
depending upon how convincing their points are. If any speaker speaks
for more than a minute, their team will get disqualified. Each member will get 15
minutes to prepare. And your time starts now!
Imagine there are two families of four people out for a ride in a
hot air balloon. Suddenly, the balloon starts moving towards the
earth instead of staying airborne. To stabilize it, one fam ily needs
to take the parachute and go out of the balloon or else, it will come
crashing down.
Who should be thrown out of the hot air balloon? *•

Reflect and Discuss:

• With the rise in Al applications replacing human workforce, do you consider it

ethical to incorporate the use of Al in various jobs?

• How do you think income would be shared if Al is used in place of human

workforce?

86 E ssentials of A rtificial I ntelligence - I X

• Al will probably bring with it many health benefits. How will these health benefits
be made accessible and available to all the people in the society?

• Al is a powerful tool in various fields. However, depending on how it is used,

it can either be a boon or a bane. Discuss.

• H ow can learning o p p o rtu n itie s fo r Al be e xtended to all?

• How will human beings ensure that they stay ahead of Artificial Intelligence?

U s e f u l T ip :
“The important thing to remember is the consequences of your actions while applying A l.”

Do you know Al can even help predict text and you can write complete stories
using Al? Maybe this book itself is written using Al! Whether we recognize it or
not, Al has already made its way into our lives. Predictive searches on Google,
‘recommended’ videos on platforms like YouTube or a software designed to prompt
and correct text while typing, are all examples of Artificial Intelligence. Generating
human-like outputs is one aspect of Al we have gained enough control over. Now,
the natural question is, can Al also make human-like decisions?

Do you know that a machine learning algorithm OpenAPs GPT - 2 language model is
trained to predict text? It takes months of training over tons of data on expensive
computers but once that is done, it is easy to use.
Recently, Al systems have been developed to analyze images to distinguish between
a benign skin lesion and melanoma cancer, matching the accuracy of 21 certified
dermatologists. This will not only make up for the lack of dermatologists but also
make their task simple and more efficient.

This example shows that Al can make good decisions, even better than humans.
One thing to notice, however, is that an algorithm is more likely to label an image
as cancer if there is a ruler in the image. Dermatologists use a ruler in the photo
to measure the size of a skin lesion. Another issue is that with algorithms like this
in place, people might not even consider seeing a dermatologist and may click a
picture of the skin lesion using mobile phones and find out if it is anything to be
worried about. However, the problem is that these photos are likely to differ from
those of a dermatologist in terms of lighting, photo quality, zoom, etc. Simple
factors like skin colour or how hairy one’s arm is can put the algorithm outside its
original training conditions. And it is well known that when algorithms are given
data that is different from
what they were trained
with, they can behave
unpredictably.

Consider the case of self­

driving cars. Imagine an
automated car is fed with
an algorithm to save the
passengers sitting in the
car over the people outside
the car, in case of accidents,
otherwise no one will buy
the car. During a collision, to
save one passenger sitting
inside the car, the car runs
over a group of six people.
This makes a self-driving
car socially unacceptable,
especially to those who
cannot afford to buy it.

88 E ssentials of A rtificial Intelligence - I X

 You can use this tool to predict text from a data set of text:

Scan QR code or visit:

http://openai.com/blog/better-Ianguage-models/

So, if Al and machines are to take over in future, how can we make their decision­
making reliable? How can we make their decisions predictable and convincing
under all circumstances? What will make these autonomous, self-improving,
independent machines and software trustworthy?

This is where ethics come into the picture. Ethics are a loosely-defined set of moral
principles about right and wrong guiding actions of an individual or a group. Since
technology itself does not possess moral or ethical qualities, it needs to be fed with
human ethics. When designed and tested well, it arrives at predictable outputs for
predictable inputs via such a set of rules or decision paths.

But here are two challenges. First, how does the team of developers determine
what is a good or right outcome and for whom? Is this outcome universally good or
is it good only for some? Is this outcome good under certain contexts or situations
and not under other conditions? Is it good against certain standards but not good
against others? These discussions, questions and answers ‘chosen’ by the team
are critical. The second challenge is that Al is an autonomous, self-learning and
self-improving technology. This means that it cannot be fed and does most of its
decision-making itself based on its own analysis of data.

SUMMARY OF WHAT ALL IS COVERED UNDER ETHICS

Basics of Al Ethics:
1. Bias, Prejudice and Fairness
2. Accountability
3. Transparency, Interpretability and Explainability

Actions of Al:

1. Safety
2. Human-AI Interaction
3. Cybersecurity and Wrong Intentional Use
4 . Data Privacy and Control

E thics in Al 89
Impact of AI:

1 . Job Losses and Unemployment

2. Civil Rights— Robot Rights
3. Human-Human Interaction Change
4 . Economy

A fundamental debate is underway that AI will change the way our society works
and it is very important to plan fo r such a society in advance before AI gets too
involved in our lives.

IF I POOP, AND YOU

SPREAD IT AROUND
THE ROOM, WHO
GETS THE BLAME?

DRAWINGSofDOGS.com

Fig. 4.2 : Blame Game in AI

At the core of these two challenges lies the problem of existing human biases which
will enter AI systems through both the developers and data. In the very process of
its creation, technology becomes inherently biased against the people who create
it. It exhibits the opinions, understanding and ethical stand of its creators. Thus,
ethics of a technology start with the ethics of its creators.

“Unfortunately, w e have biases that live in our data, and if w e don’t

acknow ledge that and if w e don’t take specific actions to address it then
w e ’re just going to continue to perpetuate them or even m ake them w orse.”
—Kathy Baxter, Ethical AI Practice Architect, Salesforce

90 E ssentials of A rtificial Intelligence - I X

 Potential Harms from Automated Decision-making
Individual Harms Collective/
Illegal Unfair Societal Harms
Loss of Opportunity
Employment Discrimination
Differential Access to Job
E.g., Filtering job candidates by race or E.g., Filtering candidates by work
Opportunities
genetic/health information proximity leads to excluding minorities
Insurance & Social Benefit Discrimination
Differential Access to
E.g., Higher termination rate for benefit E.g., Increasing auto insurance prices for
Insurance & Benefits
eligibility by religious group night-shift workers

Housing Discrimination
Differential Access to
E.g., Landlord relies on search results E.g., Matching algorithm less likely to
Housing
suggesting criminal history by race provide suitable housing for minorities
Education Discrimination
Differential Access to
E.g., Denial of opportunity for a student E.g., Presenting only ads on for-profit
Education
in a certain ability category colleges to low-income individuals

Economic Loss
Credit Discrimination
Differential Access to
E.g., Denying credit to all residents in E.g., Not presenting certain credit offers
Credit
specified neighbourhoods (redlining) to members of certain groups
Differential Pricing of Goods and Services
Differential Access to
E.g., Raising online prices based on E.g., Presenting product discounts based
Goods and Services
membership in a protected class on “ethnic affinity”
Narrowing of Choice Narrowing of Choice
E.g., P resen tin g ad s based so lely on past “clicks” for Groups
Social Detriment
Network Bubbles Filter Bubbles
E.g., Varied exposure to opportunity or E.g., Algorithms that promote only
evaluation base on “who you know” familiar news and information

Dignitary Harms Stereotype Reinforcement

E.g., Emotional distress due to bias or a E.g., Assumption that computed
decision based on incorrect data decisions are inherently unbiased

Constraints of Bias Confirmation Bias

E.g., Constrained conceptions of career E.g., All-male image search results for
prospects based on search results “CEO,” all-female results for “teacher”

Loss of Liberty
Constraints of Suspicion Increased Surveillance
E.g., Emotional, dignitary and social E.g., Use of ‘predictive policing' to police
impacts of increased surveillance minority neighbourhoods more

Individual Incarceration Disproportionate Incarceration

E.g., Use of “recidivism scores” to determine prison sentence length E.g., Incarceration of groups at higher
(legal status uncertain) rates based on historic policing data
(Source: Future of Privacy Forum)

Fig. 4.3 : Ethics-Related Issues with Al

E thics in Al 91
There are 104 cognitive biases that affect human decision-making. And, hence,
the Al we build. Al can amplify human biases by virtue of how it learns and the
feedback it gets. Such biased algorithmic systems can lead to undesirable and unfair
outcomes, unequal and unjust consequences. And the intensity of these implications
can be very severe depending on the type and number of people it affects.

Understanding bias in Al starts with tracing its sources and then identifying which
of those can be countered with technology. Bias can be of tw o types— Model bias,
which means the model is able to adequately represent the data set accurately
enough, and the other bias being Prejudice bias, where the creator’s stand on
certain issues is embedded intentionally or unintentionally in the algorithm.

“Structural bias is a social issue first and a technical issue

second.” —Kate Crawford

Let us understand this with a simple activity. Grab a sheet of paper and plot some
points on it. Now, try to draw a line through this random distribution of points.
More widely spread apart the points, curvier the line. If you continue to add more
points, the line no longer fits all the points. Now you start looking for the “ best
possible” curve fit. In Mathematics, your line can be represented by one equation
(or function) and “ the best possible” calculation by another function. This is the
fundamental idea behind machine learning model.

Fig. 4.4: Curve Fitting

List of cognitive biases:

Scan QR code or visit:
h ttps:/ / tra viswh itecommun ica tions.com/wp -con ten t/uplo ads/
2018/01 /Cognitive-Biases-Grouped.png

92 E ssentials of A rtificial Intelligence - I X

As you can see in Fig. 4.4, if you try to put the curve through maximum number
of points, you end up having a vague representation and it will be very difficult to
extrapolate and predict the nature of curve for unseen data. In fact, the equation
will be very difficult to establish fo r such a random data. Any close forced fit will
lead to underrepresentation of minority data which leads to biases. Whatever
function y=f(x) we try to establish with the data, there will always be a set of
new data fo r which the function does not hold true, so you end up making some
assumptions which are inherent biases.

Consider an example of a facial emotion recognition Al model trained on faces from

around the world. Here, we have assumed that we have enough diversity in terms
of age, gender, race, ethnicity, etc. We all know that the intensity with which we
express our emotions is also correlated with other qualitative factors. For example,
the size of smile is correlated with what is cultural appropriateness, how deeply
people relate to the trigger, how they feel about their teeth, how often they were
criticized for laughing too loudly, their mental state at that moment, etc. What
happens if we try to use this model to assess how engaged and happy students
are in a classroom? Can we have a function to represent all possible versions of
happiness on the basis of these considerations using visible and measurable features
like size of smile, volume of laughter, how wide the eyes open? What do you think?
For instance, by combining feature visualization (what
is a neuron looking for7) with attribution (how d o e s it
affect the output?), we can explore how the network
decides between labels like Labrador retriever and
tiger cat.

Several flopp y ear

detectors seem to be
im portant when
distinguishing dogs,
whereas pointy ears are
used to classify "tig e r cat*'.

CHANNELS THAT MOST

LA BR A D O R RETRIEVER
1 ■» TIGER CAT
SUPPORT ...

Fig. 4.5 : What is a neural network looking for and how is it attributing what it sees?
(Source: https://distill.pub/2018/building-blocks/)

Ethics in Al 93
As you can observe, historical bias already exists in data. Depending on how the
data set is created, Representation Bias and Measurement Bias also creep in.
Evaluation and Aggregation biases come into play depending upon the model
we deploy.

Let us now discuss accountability, safety and fairness in Al.

“There’s a real danger o f system atizing the discrim ination w e

have in society [through A l technologies]. W hat I think w e
need to do— as w e ’re m oving into this w orld full o f invisible
algorithm s every w h ere— is that w e have to be v e ry explicit, or
have a disclaim er, about w h at our error rates are like.”
—Timnit Gebru, Research Scientist, Google AI

Accountability in Al may be achieved by human audits, impact assessments or

via governance through policy or regulation. Tech companies generally prefer
self-regulation but now they realize the need for external intervention.
Governance through ‘human-in-the-loop’, where certain decisions identified as
high-risk require vetting by a human, have also been proposed as a model for
accountability.

Safety in Al can be understood as not causing accidents or exhibiting unintended

or harmful behaviour. Harm to humans is obvious and safety concerns with
autonomous vehicles, drone deliveries are well known. Did you read about the
recent attack in Saudi Arabia by an Iranian drone? Such is the power of a small
drone. How can we model and enable safety in an autonomous system?

In a rule-based engine, in which certain inputs result in a specified output, safety

measures can be put in place with testing. With Al, as the complications go far
beyond simple rule engine, it is difficult to assess all scenarios. Autonomous
decision-making requires automating the ability to evaluate safety under uncertain
conditions to predictably prevent harm.

94 E ssentials of A rtificial Intelligence - I X

 Fig. 4.6 : Containment Strategy (Source: lyad Rahwan, MIT)

Programming Al: unintended coneequences

Fig. 4.7: Unintended Consequences of Bias (Source: lyad Rahwan, m it )

“Fairness is a b ig issue. Hum an behaviour is already discrim inatory in

m any respects. The data w e ’ve accum ulated is discrim inatory. H ow can
w e use technology and A l to reduce discrim ination and increase fairness?
There are interesting w orks around ad v ersarial neural netw orks and different
technologies that w e can use to bias tow ard fairness, rather than perpetuate
the discrim ination. 1 think w e ’re in an era w here responsibility is som ething
you need to design and think about as w e ’re putting these n ew system s out
there so w e don’t have these adverse outcomes.”
—Paul Daugherty, Chief Technology and Innovation Officer, Accenture

E thics in Al 95
ANTI-BULLYING WITH AI
Bullying is something that many of us might have encountered at some point in our
life or at least read about. You should never bully anyone, help others who might
be bullied and support an anti-bullying environment in your school at all times.

It is very important to recognize when bullying is happening. Sometimes, we may

say something which we consider to be harmless but it may offend the other person
and they may feel bullied depending on their experiences, culture, religion or family
values. You should be very careful not to hurt anyone’s sentiments and beliefs.
Sometimes, you might post something on social media or send an email which
might contain some words or phrases which could be interpreted as offensive or
bullying in nature. At times, even typos, grammatical errors or sarcasm may be
misunderstood as bullying.
Can AI help?
Yes, absolutely.

Let us look at the following AI tool:

https://mycomputerbrain.net/php/experiments/ai.experiment1 9 b.php

It helps you train an AI application to recognize texts/posts and match them to

a particular output. No coding as such is required. You can enter a total of 12
expressions, including single words and sentences.

The instructions are on the right side of the menu and help you to:
• Create the training data
• Train the Artificial Neural Network (ANN) through the Start Learning button
(it will say Learning Completed when finished)
• Test the ANN by entering short sentences and phrases
On entering the inputs, ANN will be constructed in real time so you can analyze
how AI is giving the output in real life.
Note: Your input is not stored. When the browser window is refreshed, the
experiment is reset. So you can play with it as many times as you want.

Discuss your learning with your friends and come up with your own training data
to test it further.

96 E ssentials of A rtificial Intelligence - I X

 • How does Al decide when good words and when mean words are used in
combination?
• Can Al pick up sarcasm in phrases?
• Can you improve the Al model further?
19: Anti-bullying Al X-Ray Wires

Kind things £
Kind Things Mean Things
I like you You smell
Your hair looks nice 1hate you
I think you are great That's dumb
You have a great smile Your hair looks terrible
You are so clever What an idiot
I love your outfit Slimeball
Mean things

slimeball
Input Layer Hidden Layer 1 O utput Layer

This is an example table with expressions that most people would consider either
“ kind” (left-hand column) or “ mean” (right-hand column).
The grey boxes in the input, hidden and output la^er are called ‘perceptrons’. We will
learn more about perceptron model in Chapter 6 when we study Neural Networks.
A perceptron is a model or program that simulates human brain intelligence. On
the left-hand side of the ANN, we see all words and expressions in the table. They
are provided to the ANN as training data. The ANN calculates whether to classify
them into the “ Kind things” or the “ Mean things” category based upon user data.
Discuss with your friends how you can use such a tool while you are writing mails or
using social media. Did it help you understand what is considered as bullying behaviour?

E thics in Al 97
: Objective Type Questions :

A. Fill in the blanks:

1. Human biases are magnified in Al due to the way i t ................ and takes..................
2. Algorithms behave.........................when they are given data other than what they
are trained with.
3. Technology loses its neutral stand a t ...........................
4. Minority or unique features get ......................... attention in machine learning
models.
5. In a rules-based system, safety can be ensured by testing and .........................
procedures.
6. Bias can enter Al through.........................a n d ..................................
7. Ethics of technology start w ith...................................................
8. Data accumulated is ............................................... because human behaviour is
already discriminatory.

B. State whether the following statements are True or False.

1. Al can be trained to write a story from few lines of text.
2. Presenting only ads on for-profit colleges to low-income individuals is an example of
education discrimination.
3. Self-driving cars are free from any liabilities and regulations as no human is driving
the car.
4. Al biases are as good as human biases involved in algorithm.
5. All-male image search results for "CEO," all-female results for "teacher" is an example
of confirmation bias.
6. Bias in Al can be reduced with more training data.
7. Use of "predictive policing" to police minority neighbourhoods more, is inherently a
bias causing loss of liberty.
8. Al algorithms reading your emails to suggest you responses is not an example of data
protection and privacy concern.

C. Multiple Choice Questions (MCQs):

1. Machine learning understands data through:
(a) developer (b) set of codes (c) ethics (d) pattern
2. How many types of cognitive biases are there that impact human decision-making?
(a) 108 (b) 106 (c) 104 (d) 102

98 E ssentials of A rtificial Intelligence - I X

 3. Which of the following are instances of gender bias?
(a) digital assistance having female voice
(b) searching 'hands' on internet are white hands
(c) women candidate selected as 'chef' by Al system
(d) 'o bir doktor' from Turkish translated as 'he is a doctor'
4. Which of the following is untrue about Al?
(a) Al is inherently biased.
(b) Al can produce predictable output for well-designed set of rules and decision paths.
(c) Al takes up the bias of the developer.
(d) Al is unpredictable under strange situations.
5. Which one of the following is not an area of Al?
(a) computer vision (b) voice recognition
(c) web design (d) image recognition
6. Which of these does NOT use machine learning/AI?
(a) driverless cars (b) SIRI/Alexa
(c) Sonos wireless speakers (d) facial recognition on your phone
7. Which of the following is not a goal of Al?
(a) fairness (b) accountability
(c) transparency (d) obligation
8. Which algorithm is used to predict text and can be used in book writing as well?
(a) Forest Star (b) Alpha Pro
(c) Random function (d) OpenAI's GPT-2

Subjective Type Questions

1. If a machine based on Al makes a decision with unintended consequences, who is

responsible?
2. An online search shows mostly male candidates when searched for 'doctors' and
female candidates when searched for 'nurses'. Should the developer be held
accountable for this bias?
3. A company decides to select candidates using an Al system for the role of 'Head
Chef'. What all biases are likely to be present in this case?
4. In a company, visitors enter their personal details in visitors' book. However, the
company shares this data with some marketing companies. Do you think it is ethical?
Why/Why not?
5. Data privacy is a serious concern when it comes to Artificial Intelligence. What steps
are taken by various countries to deal with this problem?
6. A self-driving car is programmed to protect innocents from criminals. In case of
accidents, it hits a murderer on the road rather than a puppy. Should the developer
be held accountable for this decision?

E thics in Al 99
 7. Some experts say robots for military combats will be more efficient. However, others
challenge it on ethical grounds. What are the ethical challenges to it?
8. Human values are simple at its core. However, factors like culture, law, religion,
etc., make them complex. This complexity is at the core of ethics of Al. Discuss with
examples.
9. Do you consider it ethical to incorporate the use of Al for various jobs, given that it
takes away human jobs and magnifies human bias?
10. How do you think income would be shared if Al is used in place of human workforce?
11. Al will probably bring with it many health benefits and learning opportunities. How
will these benefits be made accessible and available to all the people in the society?
12. Al is a powerful tool in various fields; however, depending on how it is used, it can
either be a blessing or a curse. Discuss with an example.
13. How can learning opportunities for Al be extended to all?
14. How will human beings ensure that they stay ahead of Artificial Intelligence?

Imagine a major landslide occurs in a landslide-prone area in which around 20,000 people are trapped,
including pilgrims, tourists and other locals. The trapped population comprises women, children, senior
citizens, patients in hospitals, political leaders, prisoners in jail, etc. If you get to design an Al system to aid
and rescue in such scenarios, what priority order will you set for the rescue operation?

• See how Oxford University is working on an institute of Al.

Scan QR code or visit:

https:/'/www.,youtube, com /w atch ?v=5JVFKrgmr-8

What do you think about the role of such an institute? Can you think of something like this in India? Prepare
a presentation for such an institute and explain how it will work.

100 E ssentials of A rtificial I ntelligence - I X

 Understanding Al
- </>i

0 Project Cycle
(Problem Scoping, Data Acquisition, Data Exploration and Modelling)

Learning Objectives:
Ip I To understand how Al projects are designed
\p I To apply learnings to work on Al projects

B ackground a n d C ontext

“ Much has been written about A l’s potential to reflect both the best and the worst of
humanity. For example, we have seen Al providing conversation and com fort to the lonely;
we have also seen Al engaging in racial discrimination. Yet the biggest harm that Al is
likely to do to individuals in the short term is job displacement, as the amount of work
we can automate with Al is vastly bigger than before. As leaders, it is incumbent on all
of us to make sure we are building a world in which every individual has an opportunity
to thrive.”
— Andrew Ng

Introduction to Al Project Cycle

Let us go through the Al project cycle with the help of an example.

Imagine that we have come to know from our intelligence agency (it is called
Research and Analysis Wing or RAW for your reference) that there is a potential threat
to a very powerful politician (these threats are very common to political leaders)

U nderstanding Al P roject C ycle 101

and although CCTVs and basic surveillance systems are in place, they want an
Al expert like you to supervise the overall security office.

You are the Chief Security Intelligence Officer at the politician's house!

Sta rt w ith listin g do w n all th e fa c to rs w hich you need to co n sid e r w hile fra m in g
a secu rity system .

While finalizing the aim of this system, you scope the problem which you wish to
solve with the help of your project. This is Problem Scoping.

First you have to define the list of people who are allowed to enter the politician’s
house and which designated areas they are allowed to go to and for how long.
Certain key officials, VIPs, etc., may have access to any area within the premises
but this would be a defined list. The focus is to ensure that no unauthorized person
is entering the premises.

Potential steps here could be:

1. Profiles of authorized people and tag with their designated areas where they
are allowed

2. Get photographs of people who are a possible threat to the politician (this list
you can have from the intelligence agency)

3. Get photographs of all the visitors to the premises at the entry gate

4 . GPS coordinates of the locations

5. RFID tags data at the entry/exit
As you start collecting the above data, you actually acquire data in multiple
formats— coordinates, texts, photos, video, etc. This data is the source and base
of the entire Al framework, so make sure that it is accurate, reliable and traceable.
This is known as Data Acquisition.

After acquiring the required data, you realize that it is not readily usable. Some
images are low resolution, small in size, while others are better. Some images are
missing and GPS coordinates are wrong. Now you need to do Data Cleaning as
under:

• Create a database and put the names of the designated people whose profiles
you have.

102 E ssentials of A rtificial Intelligence - I X

• Put all the photographs in a model and try to derive a pattern out of it.

• Make a database to store the clean data.

• At this stage, you try to interpret some useful information out of the data you
have acquired. For this, you explore the data and try to put it uniformly for a
better understanding. This is known as Data Exploration.

After exploring the data, you now know that you need to develop a system which
detects the face of the person entering the premises and matches it with the image
data in your system to check whether this person is authorized or not. For this,
you put all your data into an Al-enabled model and train it in such a way that it
alerts the officials if an unknown person tries to enter the premises. To implement
this, you need to define the variables or parameters you need.

To implement your idea, you now look at different Al-enabled algorithms which
work on Computer Vision (since you are working on visual data). You go through
several models and select the one that matches your requirement. After choosing
the model, you implement it. This is known as the Modelling stage.

Your surveillance system is now complete! You test it by passing a mix of authorized
and unauthorized images through the model. You notice that the results are 80%
correct. After evaluating this model, you work on other shortlisted Al algorithms.
Some Al algorithms are: Minimax trees, Alpha-beta pruning, AO trees, Informed and
uninformed search algorithms like A*, Hill Climbing, Simulated Annealing, Genetic
algorithms, Optimization algorithms like anthill, etc. Do read about them!

U nderstanding Al P roject C ycle 103

As you move towards deploying your model in the real world, you test it in various
ways. The stage of testing the models is known as Evaluation. At this stage, we
evaluate each and every model tried and choose the one that gives the most
efficient and reliable results.

A fte r p ro p e r te stin g , you d ep lo y yo u r su rve illan ce syste m in th e p re m ises. A

trespasser, John, who is aware o f your smart Al-driven surveillance system, tries
to sneak in to cause damage to the system and is nabbed. Well done!

What we just did is an example of Al Project Cycle.

Let us recap the main points to understand how the entire project is executed.

Starting with Problem Scoping, you first set the goal fo r your Al project by stating
the problem you wish to resolve.

Other steps are as follows:

• You need to acquire data which will become the base of your project as it will
help you in understanding what the parameters that are related to the problem
are scoping.

• You go for data acquisition by collecting data from various reliable and authentic
sources which are traceable and auditable.

• Since the data you collect would be in large quantities, you can try to give it a
visual image of different types of representations like graphs, databases, flow
charts, maps, etc. This makes it easier for you to interpret the patterns in which
your acquired data follows.

104 E ssentials of A rtificial Intelligence - I X

• After exploring the patterns, you can decide on the type of model you would
build to achieve your goal. For this, you can research and select various models
which give a suitable output.

• You can test the selected models to figure out which is the most efficient.

• The most efficient model is now the base of your Al project and you can develop
your algorithm around it.

• Once the modelling is complete, you now need to test your model on some
newly-fetched data. The results will help you in evaluating your model and,
hence, improving it.

• Finally, after evaluation, the project cycle is complete and what you get is your
Al project.

Do It Y ourself
Now, it is your turn to describe what you have learnt. Explain the concept of Al project
cycle with the help of a suitable example and discuss with your peers in the class. Take
as an example how you could use the entire project cycle for the following project:
Let us say you are in charge o f your school lab. From the timetable, you know which
class and students need to be in the lab and at what time. How can you ensure that
only authorized students enter the lab as per their timetable?

Let us now look at each stage of the Al project cycle in greater detail.

PROBLEM SCOPING

We start with the first stage of

Al Project Cycle that is Problem
Scoping. As we have understood,
Problem Scoping means selecting
a problem which we might want
to solve using our Al knowledge.

You can select a theme from those

given below. But delve deep into
the theme to find out topics where
Fin. 5.2: A l P r o ie c t T h e m e s
problems exist.

U nderstanding Al P roject C ycle 105

After listing down the problems, go further down till the root cause of the problem
is found. We can use Ishikawa tool for this purpose.

Ishikawa or Fishbone or 5 Whys or Root-cause analysis is a tool which is intended

to reveal key relationship between various underlying variables of an effect, and
the possible causes provide additional insight into process behaviour.

The causes are plotted through analysis, brainstorming sessions, and are grouped
into categories on the main branches of the fishbone. To help structure the approach,
the categories are often selected from one of the common models shown below
but may emerge as something unique to the application in a specific case.

The beauty of this tool is that it can be applied to any situation to find out the
reasons. Let us try to apply it in the case of an organization where employees
come late. See how Ishikawa diagram can help:

Measurement

Environment I Materials
Fig. 5.3: Ishikawa diagram (So u rce : Squawkpoint)

These problems will now be very specific as they have been narrowed down from
a broader perspective. Select the root cause and make it the goal of your project.

You can either select any one out of these or you can think of one on your own.
For more options, you can also refer to the 17 Sustainable Development Goals in
Chapter 3 .

106 E ssentials of A rtificial I ntelligence - I X

For example,
In agriculture, there are pest issues, yield rates, sowing and harvesting patterns,
etc., all being very different from each other but still a part of the agriculture
theme. Thus, to effectively understand the problem and elaborate it, we need to
select one topic under the theme.

Some examples are:

Theme: Digital Literacy Topics: Online learning platforms, digital awareness,
e-books, etc. Theme: Health Topics: Medicinal Aid, Mobile Medications, Spreading
of Diseases, etc. Theme: Entertainment Topics: Media, Virtual Gaming, Interactive
AVs, Promotions, etc.

You can see that there are lots of interesting problems which can be solved using
Al. To set up the Goal of your project, you can select one problem out of the ones
listed above which you want to solve using your Al understanding. This problem
now becomes the target of your Al project and helps you get a clear vision of
what is to be achieved.

Now, you need to make the selected problem a goal. For example, a goal can
be stated as to how you can increase the production of a certain crop without
additional inputs. Even try to quantify increase in production by 20% using the
same set of inputs. Goals should always be SMART— specific, measurable, agreed
upon, realistic and time-based.

Write the Goals of your project below:

U nderstanding Al P roject C ycle 107

We will use another tool called 4 WS Problem Canvas. This canvas helps us in
identifying 4 critical parameters we need to know for solving a problem.

4Ws refer to Who, What, Where and Why.

• Who. In this stage, we look for the person who is having the problem; they are
also known as the stakeholders of the problem.
• What. Here, we consider the nature of the problem. What is the problem and
how do we know that it is a problem? Is there an anecdote or evidence to
support that it is a real problem or it is just a perception or outlier data point?
• Where does the problem arise? In this, we describe the context of the problem.
• Why refers to understanding the cause of the problem and we have already
seen how we can use Ishikawa diagram to get to the root cause of the problem.
Let us make a 4W canvas for a sample problem:

Who

The “ Who” block helps us in analyzing the people who are getting affected directly
or indirectly by it. Stakeholders are people who face this problem and would be
benefited from the solution.

Let us fill the “ Who” canvas!

Who are the stakeholders? Are these the people of the organization?
What do you know about them and their needs?

What

Under the “ What” section, you need to look into what is the actual problem
you have on hand. At this stage, you need to determine the exact nature of the
problem. What is the problem and how do you know that it is really a problem for
your stakeholders? Supporting evidence to show that this problem exists, could
be print media, TV news, government reports, etc.

Let us fill the “ What” section!

What is the exact problem? Define it.

How do you know that it is a problem? (Is there any evidence?) Show sources.

108 Essentials of A rtificial Intelligence- IX

Where

Now that you know who is affected by the problem and what the problem actually
is, you need to focus on th e c o n te xt/lo ca tio n /fre q u e n cy, etc., o f th e problem . This
section will help you look into the problem with some specificity and move you
closer to th e solution.

Let us fill th e “ W h e re ” canvas!

What is the context/situation/frequency of the problem that the stakeholders

experience?
Where is the problem exactly located?

Why

After we have figured out who is affected, know the exact nature of the problem
and where it is located, we need to understand where the solution to the problem
be deployed. Think of the benefits of this solution to the stakeholders and society
at large.

Let us fill the “ Why” canvas!

Problem Statement Template

This is a problem statement template. It is used to frame the 4WS into a consolidated
template to define the problem and its location, stakeholders and how the problem
is solved.

Problem Statement Template helps us summarize all the key points in one single
template so that in future, whenever there is a need to look back at the basis of the
problem, we can have a look at the Problem Statement Template and understand
its key elements.

U nderstanding Al P roject Cycle 109

Problem Statement Template with space to fill details according to your Goal:

[stakeholder(s)]

Our .............................................................................................................. who

[issue, problem, need]

has/have .............................................................................................
a problem ............................................................................................. what
that

[context, situation]

when/
while where

[benefit of solution for them]

An ideal
solution why
would

(Source: CBSE)

DATA ACQUISITION
We have learnt how to scope a problem and set a goal for the project. After setting
the goal, we list down all the necessary elements which are directly/indirectly
related to our problem. This is done using the 4Ws problem canvas. The 4Ws are:

1. Who
1. Who are the stakeholders?
2. What do we know about them?
2 . What
1. What is the problem?
2. How do we know that it is a problem? (Is there an evidence?)
3 . Where
1. What is the context/situation/frequency of the problem?
2. Where is the problem located?

110 Essentials of A rtificial Intelligence- I X

4 . Why

1. What would be beneficial for the stakeholders?

2 . How will the solution improve their situation?
3 . To summarize, we go fo r the problem statement template where we put in
all the details.

Our ....................... (Stakeholders) ....................... has/have a problem

that ....................... (issue, problem, need) ....................... when/while
.......................(context, situation). An ideal situation would b e ........................
(benefit of solution for th e m )........................ We can ensure that this solution
is working reliably th ro u g h ..........................

As we move ahead in the Al Project Cycle, we come across the second element
which is called Data Acquisition. Data can be a piece of information or facts
collected together for reference or analysis. Whenever we want an Al project to
be able to predict an output, we need to train it first using data-training data set.

For example, suppose you want to make an Al-powered system which, based on
historical data, can predict whether an employee will come late the next day or
not. This is the data with which the machine can be trained. Once it is ready, it will
efficiently predict whether the employee will come late or not. The previous salary
data here is known as historical data or training data while the next late-coming
prediction data set is known as the testing data.

For better efficiency of an Al project,

the training data needs to be relevant and
authentic. In the above example, if the
training data was not of his late coming
but of salary, the machine would not
have predicted his next late-coming event
correctly since the whole training would
have gone wrong. Similarly, if the previous
attendance was not authentic, that is, if it
was not correct, then also the prediction
could have gone wrong, (source: c b s e )

U nderstanding Al P roject Cycle 111

DATA FEATURES
For any Al project to be efficient, the training data
should be authentic and relevant to the problem
statement scoped.

We have now come to the stage of data acquisition.

How do we know what data to get based on the
problem statement? We need to visualize the
factors which affect the problem statement. For
this, we need to extract Data Features for the
problem scoped. Now try to find out what are the
parameters which affect your problem statement
directly or indirectly and list them down.
Look at your problem statement once again and try to find out the data features
required to address this issue. Data features refer to the type of data you want
to collect. In our previous example, data features would be day, in time, out time,
biometric machine attendance data, leave record, etc.

Try to identify the data features required for your problem statement:

Acquiring Data from Reliable Sources

In a box, write problem statement with data features arrows.

It is very important to find reliable data as data is the foundation of all Al projects.
After finding out the data features, we now need to acquire the same. There exist
various sources from which data can be acquired. Do all the sources have authentic
data? What if we do not get appropriate data?

112 Essentials of A rtificial I ntelligence- IX

Data plays an important role in an Al project as it creates the base on which the
Al project is built. Therefore, the acquired data should be authentic, reliable and
correct. Also, the acquisition methods should be authentic so that our project is
not in conflict with anyone.

After mentioning the data features, we get to know what sort of data is to be
collected. Now, the question arises: From where can we get this data? There can
be various ways in which we can collect data. Some of them are:
• Surveys • Web Scraping • Sensors
• Cameras • Observations • APIs

Sometimes we use the internet and try to acquire data for our project from some
random website. Such data might not be authentic as its accuracy cannot be proved
unless proper sources are mentioned. Thus, it becomes necessary to find a reliable
source from where authentic information can be taken. Extracting private data can
be an offence. Europe has a strict law against it called General Data Protection Rights
(GDPR). One of the most reliable and authentic sources of information is the open-
sourced websites hosted by the government. These government portals have general
information collected in suitable format which can be downloaded and used wisely.

Some of the open-sourced government portals are: data.gov.in, india.gov.in, etc.

List down ways of acquiring data for a project below:

SYSTEM MAPS
A System Map helps us in finding relationships between different elements of
the problem which we have scoped. It helps us in strategizing the solution for
achieving the goal of our project. Here is an example of a system very familiar to

U nderstanding Al P roject Cycle 113

you—Water Cycle. The major elements of this system are mentioned here. Look at
these elements and try to understand the system map for this system. Also, look
at the relation between all the elements. After this, make your own system map
for the data features which you have listed. You can also use the online animated
tool fo r creating your system maps.

Resources
You can make system maps online using an animated tool:
Scan QR code or visit:
https://ncase.me/loopy/

We use system maps to understand complex issues with multiple factors that affect
each other. In a system, every element is interconnected. In a system map, we try
to represent that relationship using arrows. Within a system map, we identify loops.
These loops are important because they represent a specific chain of causes and
effects. A system typically has several chains of causes and effects. You may notice
that some arrows are longer than the others. A longer arrow represents a longer
time for a change to happen. We also call this time delay. To change the outcome
of a system, as a change-maker, we have tw o options— change the elements in a
system or change the relationship between elements. It is usually more effective to
change the relationship between elements. You may also notice the use of V signs
and signs. These are an indicator of the nature of relationship between elements.

A system map shows the components and boundaries of a system and the
components of the environment at a specific point of time. With the help of system
maps, one can easily define a relationship between different elements which come
under a system. The goal of our project becomes a system whose elements are
the data features mentioned above. Any change in these elements changes the
system outcome too.

Here is a sample System Map:

THE WATER CYCLE

The concept of water cycle is very simple to understand and is known to us all.
Water is in rivers; it is steam and in clouds as well. It explains how water completes
its cycle transforming from one form to another (liquid, solid and gas).

114 Essentials of A rtificial Intelligence- I X

The elements which define the water cycle system are:

Snow Underground Soil Rivers

Oceans Trees Land Animals

Fig. 5.4: Water Cycle System

Let us draw the System Map for the Water Cycle now.

In this system map, all the elements of the water cycle are put in circles. The map
here shows cause and effect relationship of elements with each other with the
help of arrows. The arrow head depicts the direction of the effect and the sign
(+ or -) shows their relationship. If the arrow goes from X to Y with a + sign, it
means that both are directly related to each other. That is, if X increases, Y also
increases and vice versa. On the other hand, if the arrow goes from X to Y with a
- sign, it means that both the elements are inversely related to each other, which
means if X increases, Y will decrease and vice versa.

U nderstanding Al Project Cycle 115

Now, it is your turn to create your own System Map!
Considering the data features for your problem, draw a system map in the box
provided.

D A T A E X P L O R A T IO N

To explore various data visualization techniques, visit

Scan QR code or visit:

h ttps://datavizca talogue. com/

On this website, you will find various types of graphical representations,

flow charts, hierarchies, process descriptors, etc. Go through the page and
look at various new ways and identify the ones which interest you the
most.

Why do you think we need to explore and visualize data before jumping to the Al
model? When we pick up a library book, we tend to look at the book cover, read
the back cover and skim through the content of the book prior to choosing it as
it helps us understand if this book is appropriate for our needs and interests.

Similarly, when we get a data set in our hands, spending time to explore it will
help us get a sense of the trends, relationships and patterns present in the data.
It will also help us better decide which model/models to use in the subsequent
Al Project Cycle stage. We use data visualization as a method because it is much
easier to comprehend information quickly and communicate the story to others.

In the previous modules, you set the goal of your project and also found ways to
acquire data. While acquiring data, you must have noticed that the data consists
of numbers, words, media, etc. For example, if you go to a library and pick up a

116 E ssentials of A rtificial Intelligence - I X

random book, you first try to go through its contents quickly by turning pages and
by reading the description before borrowing it for yourself because it helps you in
understanding if the book is appropriate for your needs and interests you or not.
To analyze the data, you need to visualize it in some user-friendly format so that
you can:
• Quickly get a sense of the trends, relationships and patterns contained within
the data.
• Define strategy for which model to use at a later stage.
• Communicate the same to others effectively. To visualize data, we can use
various types of visual representations.

Are you aware of visual representations of data? Fill them below:

m r n m t ------------------------------------------------------------------------
^L ist down 5 new data visualization techniques which you have learnt
Scan QR code or visit:
https://datavizcatalogue.com/

1 Data Visualization Technique 1 Data Visualization Technique 2 1

Name of the Representation Name of the Representation
One-line Description One-line Description
How to draw it How to draw it
Suitable for which data type Suitable for which data type

SKETCHY GRAPHS
In this activity, we are going to sketch graphs! Now that you have explored various
types of graphs and have already chosen the best ones to plot your data features,
do this activity in groups. Select any tw o data features and plot their graphs on a
chart paper. Make sure that you are able to relate this graph to the goal of your
project.

Let us now look at the scoped problem statement and the data features identified
for achieving the goal of your project. Try looking for data required for your project
from reliable and authentic sources. If you are not able to find data online, try
using other methods of acquiring data (as discussed in the Data Acquisition section
earlier).

U nderstanding Al P roject C ycle 117

Once you have acquired the required data, you need to visualize it. Under the
Sketchy Graphs activity, visualize your collected data in a graphical form at for
better understanding. For this, select one of the representations from the link or
choose the one which you already know about. The basis of your selection should
be the data feature which you want to visualize in that particular representation.
Do this for all the data features that you have for the problem you have scoped.
Let us now answer the following questions for a better understanding:

Let us start drawing visual representations for all the data features extracted and
try to find a pattern or a trend from it.
For example, if the problem statement is: How can we predict whether a song will
make it to the top 10 in a famous countdown?

We would require data features like: Current trends of music, genre, style, duration
of song, popularity of the movie or singer, etc.
KPI Dashboard .• M Z . ■»•).
§1)
New Paoe

Custom er Satisfaction “^ ^ ^ ^ rOSSProfit New Custom ers Revenue

9 3 .1 3 % $ 1 9 2 .1 3 M 1 0 ,7 1 9 $ 1 ,0 6 1 M
Previous
79.82
%Change
+14.29
Trend Previous
$183.61M
%Change
+4 43%
Trend Previous
11,918
% Change
-11.21%
Trend
▼
Previous
S906M
%Change
+14.61%
Trend
^

Sales by Product Category Sales Comparison Sales by Month Brand Profitability

■ Accessories
H Camcorder
■ Computers
a
■
■
Media Player
Stereo Systems
Televisions IH IIH Ii H 1
■ Video Production
■ This Year ■ Last Year
szsasszssSiS
Sales by Country P ro d u c ts
Sales by Count]y ▼ Stores Customers Comments

Subcategory Gross Profit Discount MSRP COGS Qty

| Blu Ray S51.771.195 | M M S10.895.633 243.779,705 S181.112.921 679,495

ill • Speaker Kits

Headphones
S25.819.242

S24.523.024
■

warn
■ $4,954.243

S3,516,913
112,169.618

79.703,501
S81.396.140

S51,663.564
244,199

228.349

Handheld S21,393,655 mm SI,959,624 43,930,192 S20.576.916 250,167

Standard S19.369.668 wm S3,214,787 71,656,083 S49.071.633 192,205

Video Editing S17,947.620 ■ i $2,695.891 60,749,162 $40,105,657 199,749

Tablet S17,674,116 H S2.018.135 45.464.132 S25.771.890 146,728 |

Receivers $16,555,836 ■ 1 $2,643,045 59,528,536 S40,329,668 150,568

Rat Panel TV SI 5.885.499 m S3,478,829 78,441,670 S59,077,345 92,501

Smartphone S15.834.702 m S2.790.776 62,661.241 S44.035.774 205.049

j Professional $8,835.523 ■ SI ,933,997 45.987.828 S35.218.308 12,872 i

i****"- I,
| 2000 mi ;
Charger SI,970,124 i $187,486 4,210,324 S2.052.711 105.257

Streaming SI.936.587 i $338,560 7,339.881 S5.064.730 67.910

2 Comments

(Source: W e b F O C U S )

Now, to analyze a pattern, we can say that the popularity of the actor will have
a direct bearing on the output of the system. So, we will plot a graph showing
the popularity of various singers and the one who is most popular and has the

118 E ssentials of A rtificial Intelligence - I X

maximum chance of getting to the billboard. This way, the graphical representation
helps us understand the trends and patterns from the data collected and to design
a strategy around them for achieving the goal of the project, (source: c b s e )

pBsjjjr
r Take a chart paper and start representing your data features in various types of
graphs. After completing this exercise, present your work to your friends in the class
and explain to them the trends and patterns that you have observed.
Now list down the trends that you have observed in your representation.
v____________________________________ _____________________________________

M o d e l l in g

In the previous module of Data Exploration, you explored the data you had
acquired at the Data Acquisition stage for the problem you scoped at the Problem
Scoping stage. Now you have visualized some trends and patterns out of the data
which would help you develop a strategy for your project. To build an Al-based
project, you need to work around models or algorithms to predict the output
given a set of inputs. This could be done either by designing your own model or
by using the pre-existing Al models. Before jumping to modelling, let us clarify
the definitions of Artificial Intelligence (Al), Machine Learning (ML) and Deep
Learning (DL):

• Artificial Intelligence or Al refers to any technique that enables computers to

mimic human intelligence. An artificially intelligent machine works on algorithms
and data fed into it and gives the desired output.

• Machine Learning or ML enables machines to improve upon tasks with

experience. The machine here learns from the new data fed into it while testing
and uses it for the next iteration. It also takes into account the times when it
went wrong and considers the exceptions too.

U nderstanding Al P roject C ycle 119

• Deep Learning or DL enables the software to train itself to perform tasks with
a vast amount of data. Since the system has got a huge set of data, it is able
to train itself with the help of multiple machine learning algorithms working
together to perform a specific task.

Artificial Intelligence is the umbrella term which holds both Deep Learning as well
as Machine Learning. Deep Learning, on the other hand, is a very specific learning
approach which is a subset of Machine Learning as it comprises multiple Machine
Learning algorithms.

As you have been progressing towards building Al readiness, you must have come
across a very common dilemma between Al and ML. Many a time, these terms
are used interchangeably but are they the same? Is there no difference between
Machine Learning and Artificial Intelligence? Is Deep Learning also Artificial
Intelligence? What exactly is Deep Learning?

ARTIFICIAL
INTELLIGENCE
A program that can sense, reason,
act and adapt

MACHINE LEARNING
Algorithms whose performance
improves as they are exposed to
more data over time

DEEP
LEARNING
Subset of machine
learning in which
multilayered neural
networks learn from
vast amounts
.. of data

As you can see in the above diagram, Artificial Intelligence is the umbrella term
which covers both machine and deep learning. Deep Learning comes under Machine
Learning. It is a funnel-type approach where there are a lot of applications of Al, out
of which a few are those which come under ML and out of those, very few go into DL.
Al Modelling refers to developing algorithms which can be trained to get intelligent
outputs which we can define as y=f(x).
We have seen various types of graphical representations and data visualization
techniques which can be used for representing different parameters of data.
The graphical representation makes the data understandable for humans as we
can discover trends and patterns in it. But when it comes to machine accessing
and analyzing data, it needs the data in the most basic form of numbers

120 E ssentials of A rtificial Intelligence - I X

(which is binary— Os and Is) and when it comes to discovering patterns and trends
in data, the machine goes for mathematical representations of the same. The
ability to mathematically describe the relationship between parameters is the heart
of every Al model. Thus, whenever we talk about developing Al models, it is the
mathematical approach towards analyzing data which we refer to. (so u rc e : c b s e )

Fig. 5. 6: Rule-based Al Approach

Rule-based Approach refers to Al

Modelling where the relationship or
patterns in data are pre-defined. The
algorithm just follows the rules or
instructions mentioned and performs
its tasks accordingly. For example,
suppose you have a data set comprising
1,000 images of onions and as many
of carrots. To train your machine,
you feed this data into the machine
and label each image as either onion
or carrot. This is your training data.
Now if you test the machine with the
image (Testing Data) of an onion, it will compare the image with the trained data
and as per the rules, it will identify the test image as an onion. This is known as
Rule-based Approach.
Learning-based Approach refers to Al Modelling where the underlying relationship
or patterns in the data are not defined. In this approach, random data is fed into
the machine and the algorithm needs to derive a relationship in the data. Generally
this approach is followed when the data is unlabelled or too random fo r a human
to make sense out of it. The machine analyzes the data and tries to extract similar
features out of it and clusters same data sets together. In the end as output, the
machine gives us broad trends observed in the data set.

U nderstanding Al P roject C ycle 121

For example, imagine you have a data set of 10,000 images of people in your city
and you have to understand which of them are sick. Obviously, you would not have
any ready reference as to how to identify from a picture itself that someone is not
well. You might use their facial expressions and other emotions to cluster them into
groups and try to understand what attributes in the image define a sick person.

Think how you could do it. Do it as a fun activity.

DECISION TREE
Do you remember the Story Speaker activity we did in Chapter 2? We had set up
certain conditions and outcomes to guide our friends around our smart home.
Decision trees are similar to that. They are an example of a rule-based approach.
The basic structure of a decision tree starts from the root. From there, the tree
diverges into multiple directions with the help of arrows called branches. These
branches depict the condition because of which the tree diverges. In the end, the
final decision is where the tree ends. These decisions are termed as the leaves of the
tree. If you look carefully, you will realize that this looks like an upside-down tree.

While making a decision tree, ask the students to look at the data carefully. Ask
them to figure out which out of the whole data would be the root and the leaves.
After this, they should analyze the data and find out if there is some redundant
data which might not be necessary while making the decision tree.

In the Chapter, you developed an interactive story using the Google Docs
extension called Story S p e a k e r . W rite the story below:

122 E ssentials of A rtificial I ntelligence - I X

Decision trees are similar to the concept of Story Speaker. It is a rule-based Al model
which helps the machine in predicting what an element is with the help of various
decisions or rules fed into it. The basic structure of a decision tree is shown below:

In the above example, the decision tree starts with the question ‘Am I hungry?’
The starting point of any decision tree is known as its Root. It then branches into
tw o different ways or conditions: Yes or No. The forks or diversions are known as
branches of the tree. The branches either lead to another question, or they lead
to a decision like ‘Go to sleep’ which is known as the leaf. If you look closely at
the image above, you will notice that it looks like an inverted tree with root above
and the leaves below. That’s why it is called a Decision Tree!
Decision trees are made on the basis of the data set we have and change according
to the parameters which we take into consideration for modelling. Many a time, the
data set might have redundant data, that is, some data might not hold importance
while developing a decision tree. For this, one needs to visualize the relation
between all the parameters given in the data and then formulate the model.

Things to Remember
• Ensure that you understand the data set properly before you start making a Decision
Tree. Try selecting any one output and, on its basis, find out the common links
which all the sim ilar outputs have and create branches.
• Many a time, the data set m ight contain redundant data which does not hold
any value while creating a decision tree. It is important to note which are the
parameters that affect the output directly and only those should be used while
creating a decision tree.
• There m ight be multiple decision trees which lead to correct prediction for a single
data set. The one which is the simplest is considered the best.

U nderstanding Al P roject C ycle 123

 Let us do a simple activity to understand how a decision tree is made.
The following is a data set comprising 4 parameters which lead to the prediction
whether an elephant would be spotted or not. The parameters which affect the
prediction are: Weather Outlook, Temperature, Humidity and Wind. Draw a decision
tree for this data set.
Outlook Temperature Humidity Wind El
S|
Sunny Hot High Weak No
Sunny Hot High Strong No
Overcast Hot High Weak Yes
Rain Mild High Weak Yes
Rain Cool Normal Weak Yes
Rain Cool Normal Strong No
Overcast Cool Normal Strong Yes
Sunny Mild High Weak No
Sunny Cool Normal Weak Yes
Rain Mild Normal Weak Yes
Sunny Mild Normal Strong Yes
Overcast Mild High Strong Yes
Overcast Hot Normal Weak Yes
Rain Mild High Strong No
(Source: C B S E ) /k , kj ; l
V/ h A
Draw your decision tree here.

V )

124 E ssentials of A rtificial I ntelligence - I X

PIXEL IT!

As we have discussed earlier, there are different kinds of Al models available. We

discussed a rule-based approach Al model called Decision Tree. We will now explore
other types of Al Modelling. We will learn how the computer classifies images and
how the computer processes them.

This is an example of machine learning approach that is typically used in Computer

Vision applications.

Let us start with an activity. Follow the instructions given below:

• Cut out the matrix from the page given below or draw the same on a blank page
with 6x6 square blocks.
• Write an uppercase alphabet on this matrix. The height of the alphabet should
be equal to the height of this matrix. In other words, it should start from the
bottom line of the matrix till the top line. You can write any capital alphabet in
any handwriting.
• Now, colour the boxes on which the lines of that alphabet have fallen.
• After this, cut out horizontal stripes of the matrix such that it goes from 1-2 , 2 -3 ,
3 -4 , 4 -5 , 5 -6 and 6 -7 .
• Now, paste all these stripes together to form a single paper string. Make sure
that the last block is neither over the first block of next line nor is there any gap
between the first and the last blocks.
• Now, find students in your class who have chosen the same alphabet as you. Put
their paper strings under your string and add up all the coloured blocks to get a
series of numbers. A block without colour counts as 0 while the coloured ones
count as 1. If a column has 3 coloured boxes, the summation turns out to be 3.
• Now, get another student whose letter is different from yours. Put his paper string
under your multiple strings (of same alphabet) and see if the pattern of coloured
blocks is same or not.
• Also, go to other groups and check if their summation series of numbers is the
same as yours or not.
• Note down your observations.

U nderstanding Al P roject C ycle 125

 Cut out the matrix

1 2

2 3

3 4

4 5

5 6

6 7

Write an uppercase alphabet on this matrix. The height of the alphabet should be
equal to the height of this matrix. In other words, it should start from the bottom line
of the matrix till the top line. You can write any capital alphabet in any handwriting.

126 E ssentials of A rtificial Intelligence - I X

 Colour the boxes on which the lines of the alphabet have fallen.

Cut out horizontal stripes of the matrix such that it goes from 1-2, 2-3, 3-4, 4-5,
5-6 and 6-7.
1 2 3

Paste all these stripes together to form a single paper string. Make sure that the
last block is neither over the first block of next line nor is there any gap between
the first and the last blocks.

L m 11 i i
0341001430

U nderstanding Al P roject C ycle 127

 Find students in your class who have chosen the same alphabet as you. Put their
strings under your string and do addition of all the coloured blocks to get a series
of numbers. Blocks without colour count as 0 while the coloured ones count as 1.
If a column has 3 coloured boxes, the summation turns out to be 3.
Follow the rest of the steps to finish this activity.
v------------------------------------------------- -------------------------------------------------------------- ;

What we have learnt

Pixel It! activity is an example of how computers see images, pre-process, process
and classify them. This kind of Machine Learning approach is commonly used in
Computer Vision-related applications which we learnt in Chapter 2. Every image
which is fed into the computer is divided into pixels (smallest unit of an image).
The computer analyzes each pixel and if it has to compare tw o pictures to check
if they are similar or not, a pixel-wise comparison takes place. If all the pixels are
identical, this means that the images are exactly the same.

In this activity, you created your own intelligent model which can identify if an
alphabet is similar to the data trained or not. Here, you divided your alphabet into
36 blocks and then processed it. Also, to train the model, you used data sets of
the same alphabet but in different handwritings. Now, when you go for testing
it, the model would see if the coloured blocks are aligned or not. If majority of
the blocks are aligned, there exists maximum probability that the alphabet is the
same. Else, the alphabet is different, ■
(so u rc e . c b s e )

In the Decision Tree activity, we tried a rule-based approach and in the Pixel It!
activity, we tried the machine learning approach. In machine learning, what you
want to accomplish is for the machine to produce the model for you. What you
will provide is the training data and the machine will undergo a training process
and produce the model.

Some of the common Al models are given below:

Regression: This is a type of rule-based Al model in which the algorithm generates

a mapping function from the given data, represented by a line. Data is plotted as
shown in the graph below and the solid line here represents the mapping function
for them. With the help of this mapping function, one can predict and forecast the
future data. For example, if we want to predict what will be your marks in the next

128 E ssentials of A rtificial Intelligence - I X

exam in a particular subject, we can use historical data and analysis to forecast.
Regression works with continuous data.

Classification: This is another rule-based Al modelling technique. In classification,

the algorithm is able to determine which points in the data set belong to either side
of the classification function represented by the dotted line. The model classifies
data sets according to the rules given to it as inputs. Usually the data set used for
classification is labelled data and is sorted as per labelling done.

For example, if we want to train a model to identify if an image is that of a man

or woman, we need to train it with multiple images of both men and women in
different structures and formats. The machine will then classify images on the
basis of training data set. Classification works on discrete data set.

Clustering: This is a machine learning model where the machine generates its own
rules or algorithms unlike a rule-based model. The data fed into such a model is
usually unlabelled or random.

We need to devise our algorithms in a way that they are generated on their own
based on the data set. Machine needs to derive patterns or trends from the
training data set to cluster the ones which follow the same pattern (See Fig. 5.10).
For example, if you have a random data of students in your school, and you are
unable to find out to which class or age group they belong, you will feed their
data into the clustering algorithm. The algorithm would then analyze the data and
divide them into clusters according to their similarities of variables or behaviours
exhibited and derive patterns based on the trends noticed. Clustering works on
discrete data set.

U nderstanding Al P roject C ycle 129

 Fig. 5.9: Clustering and Classification

EVALUATION
After you have completed all the cycles of the Al project cycle, it is very important
to keep evaluating your model to ensure that the results are working well with
the new data. If there are variances between the training data set and test data
set, keep iterating on the model to improve.

: Objective Type Questions

A. Fill in the blanks:

1 ............................. is a Machine Learning model where the machine generates its own
rules or algorithms unlike a rule-based model.
2 ............................. as a method helps to comprehend information quickly and
communicate the story to others.
3 . A ............................ shows the components and boundaries of a system and the
components of the environment at a specific point in time.
4 . Data for an Al project can be collected through............................, ............................ and

5 . In classification, the algorithm is able to determine which points in the data set belong
to either side of the classification function represented by t h e ..............................
6. Al algorithms are trained on data which is called ..............................
7 . Data acquired in the project should be ............................, ............................ and

8................................is used to frame the 4 Ws into a consolidated template to define the

problem and its location, stakeholders and how the problem is solved.

B. State whether the following statements are True or False.

1 . Al can be trained to write a story from a few lines of text.
2 . Classification works on discrete data set.

130 E ssentials of A rtificial Intelligence - I X

 3 . Data fed into a Clustering model is usually labelled.
4 . Decision Trees are an example of rule-based approach.
5 . System Maps help us to find relationships between different elements of the problem
which we have scoped.
6. Regression is a rule-based Al model.
7 . Data features refer to the type of data you want to collect.
8. Learning-based approach refers to Al modelling where the underlying relationship or
patterns in the data are defined.

C. Multiple Choice Questions (MCQs):

1 . Regression only works with:
(a) Intermittent data (b) Continuous data
(c) Step function data (d) Partially linear data
2 . Smallest unit of an image which the computer uses in CV is called:
(a) KB (b) Data (c) Pixel (d) Byte
3 . Which of these means the type of data needed for an Al project?
(a) Regression (b) Data Features
(c) Classification (d) Clustering
4 . The basic structure of a Decision Tree starts from:
(a) Root (b) Branch (c) Leaves (d) Stem
5 . Which of these is not a visualization technique in Al Project Life Cycle?
(a) System Maps (b) Neural Network
(c) Sketchy Graphs (d) Decision Tree
6. Which is the first stage of Al project life cycle?
(a) Evaluation (b) Problem Scoping
(c) Data Acquisition (d) Modelling
7 . What are the people facing a problem called?
(a) Stakeholders (b) Activators
(c) Influencers (d) Actors
8. Which of the following is not a W in 4 W framework?
(a) Where (b) Which (c) Who (d) What

: Subjective Type Questions !

1 . Explain how you think system maps are useful in defining the workflow in Al projects.
2 . Do you think Data visualization is important in defining Al projects? If yes, explain
how.
3 . What are the key differences between the Rule-based and Learning-based Al?

U nderstanding Al P roject C ycle 131

 4 . At what stage of Al project cycle should you take care of biases and why?
5 . What are the key differences between Classification and Clustering techniques? Take
an example of a sample data set and plot graphs to explain.
6. Apply Al project cycle to come up with an innovative solution of managing library
books in your school—the whole process of search, issuing and returns management.
7 . Explain the importance of Decision Trees and give examples of how they can be used
in assessments in learning.
8. What is the importance of data features in a project? Taking the example of predicting
the weather, what all data features will you use?

Quick Activity
i
V W w w w w
Check out Harvard's article on how to make Al projects successful:

Scan QR code or visit:

https://hbr.org/2018/07/how-to-make-an-ai-project-more-likely-to-succeed

List down your learnings from this and explain how you will make your Al projects done in the Chapter
taking care of these learnings.

Read how Al devices work and the thought process behind the design. This is how you should think when
the end result of your Al project is a product or device.

Scan QR code or visit:

https://anatomyof.ai

132 E ssentials of A rtificial I ntelligence - I X

 Getting Started with
Neural Networks
Learning Objectives:
Ip [ Introduce the basics of neural networks
Ip I Understand how neural networks work in the context of Al

B ac kg r o u n d a n d C ontext
“ I think the brain is essentially a computer and consciousness is like a computer
program. It will cease to run when the computer is turned off. Theoretically, it could be
recreated on a neural network, but that would be very difficult, as it would require all
one’s memories.”
— Stephen Hawking

Do you have an idea how our brain works? How

we detect things, how we recall a memory or
how we think? We have a network of neurons
in our brain which fire together to make
connections. It is generally said “ Neurons that
fire together, wire together” , wire being the
connection.

Neural networks are loosely modelled after

how neurons in the human brain behave.
The key advantage of neural networks is that
they are able_to_extract data features automatically without needing th e jn p u t
of the programmer. A neural network is a system of organizing machine learning
algorithms to perform certain tasks. It is a fast and efficient way to solve problems
for which the data set is very large (see image on the next page).

G etting S tarted with Neural Networks 133

This is a representation of how neural networks work. A Neural Network is divided
into multiple layers and each layer is further divided into several blocks called
nodes. Each node has its own task to accomplish which is then passed on to the
next layer. The first layer of a Neural Network is known as th e input layer. The job
of an input layer is to acquire data and feed it to the Neural Network. No processing
occurs at th e input layer. Next to it are the hidden layers. Hidden layers are those
layers where the whole processing occurs. Their name essentially means that these
layers are hidden and not visible to the user.

Each node of these hidden layers has its own machine learning algorithm which it
executes on the data received from the input layer. The processed output is then
fed into the subsequent hidden layer of the network. There can be multiple hidden
layers in a neural network system and their number depends on the complexity
of the function for which the network has been configured. Also, the number of
nodes in each layer can vary accordingly. The last hidden layer passes the final
processed data to the output layer, which then gives it to the user as the final
output. Similar to the input layer, output layer too does not process data which
it acquires. It is meant for user interface.

“Your brain does not m anufacture thoughts. Your thoughts shape

neural netw orks.”
—Deepak Chopra

134 E ssentials of A rtificial Intelligence- I X

NEURAL NETWORKS Vs HUMAN NERVOUS SYSTEM
Let us recall how neurons in a human body work.

impulses carried
towards cell body
hranrhps

Dendrites Axon
terminals

Nucleus

Cell Body
Fig. 6.2 : How Neurons Work

Given above are the images of a human neuron and its relation with the Neural
Network. The axon from a neuron sends an impulse to the synapse of another
neuron. The impulse received is then sent to the cell body (nucleus) through
dendrites. The cell body performs an activation function on the impulse received
and then gives it to the output axon which passes the same to the next neuron in
the system. Now, as we relate this process with an Artificial Neural Network, we
can see that the input layer gets data which is passed on to the nodes in the hidden
layer. The nodes perform specific actions on the data and pass the processed
information to the next layer. In the end, the final processed data reaches the
output of the system.

Our brain uses an extremely large interconnected network of neurons

fo r processing information and to model the world around us. A neuron
collects inputs from all other neurons using dendrites and then it sums up all
the inputs and if the resulting value is greater than a defined threshold,
it fires. This fired signal is then sent to other connected neurons through the
axon.

G etting S tarted with Neural Networks 135

BUILDING A FUNCTIONAL MODEL OF BRAIN

Conventional HOOVER DAM

hardware Produces up to
MEGAWATTS 2 gigawatts
Neuromorphic SUPERCOMPUTER
hardware CHIP
Not analoguous to neural
function. Runs slower
than biological neural
networks
WIND
TURBINE
Produces ~ 2-3
megawatts

20 W atts
BRAIN 5 0 -1 0 0 WATTS
~85 billion neurons DESKTOP
1 quadrillion synapses COMPUTER
CHIP
Runs vastly slower
SPINNAKER than biological
Variable num ber of neural networks
artificial neurons,
LIGHT BULB
typically ~ 1 ,000
Uses ~20 watts
1 million synapses per
1,000 neurons
Runs at speed of biological
neural networks
' ■ BRAINSCALES
,WATT
5 12 artificial neurons
128,000 synapses
LASER IN 10,000x faster than biological neural networks
CD/DVD
PLAYER
TENS OF MILLIWATTS
Uses
-5 -1 0 milliwatts LOIHI
-13 0 ,0 0 0 artificial neurons
130 m illion synapses
Runs at speed of biological neural networks

TENS OF MILLIWATTS
IBM'S TRUE NORTH
~ 1 million artificial neurons
256 m illion synapses
Runs at speed of biological neural networks
| HEARING AID
Uses less than 1 milliwatt

Fig. 6.3: Power requirements for building a brain-like device

(Source: Unisem Europe Ltd; © Heidelberg University, Intel Corporation, IBM Research)
wVJu. ^ ^ ft
w h at are A r t if ic ia l neural netw orks (A n n s )
Artificial Neural Networks (ANNs) are essentially algorithmic representations of
our brain’s neuronal structure. As we have learnt earlier, neurons connect with
each other and make connections which gradually lead to a complicated network
structure. You should note that the output of any given neuron may be the input
to thousands of other neurons, that’s how complicated it can get! Learning occurs
by repeatedly activating certain specific neural connections over others, and this

136 E ssentials of A rtificial Intelligence- I X

reinforces those connections. This leads to memory and habits because it makes
neurons more likely to produce a desired outcome given a specified input. This is
actually called feedback— when the desired outcome occurs, the neural connections
causing that outcome become stronger.
ANNs try to mimic this same behaviour of our brain and they can be trained in
a supervised or unsupervised manner. In case of a supervised ANN, the network
is trained by providing matched input and output data set, with the intention of
getting ANN to provide a desired output for a given input (Y=f(x)).
An example is an email spam filte r algorithm—the input training data could be
the type of words in email such as “ free” , count of various words in the body of
the email, subject of the mail and the output training data would be a classifier
of whether the email was truly spam or not. If lots of emails are passed through
the neural network as training data set, this allows the network to learn what
input data makes it likely that an email is spam or not. This learning takes place
by adjusting the weights of the ANN connections which we will discuss shortly.

“The neural network is this kind of technology that is not an

algorithm, it is a network that has weights on it, and you can
adjust the weights so that it learns. You teach it through trials.”
—Howard Rheingold

The above figure depicts a neuron connected with n other neurons and receives
n inputs (xlf x2, ..... , xn). This standard configuration is known as a Perceptron.

It contains the inputs {xv x2, ....., xn) and weights w2, , wn) which are real
numbers and can be positive or negative.

G etting S tarted with Neural Networks 137

A perceptron consists of weights, summation processor and an activation function.
It also contains a threshold processor, commonly known as bias, to define the
threshold level. Remember, only when threshold level is reached is a neuron fired.

All inputs are individually weighted, then added together and passed onto the
activation function. There are many types of activation functions but the simplest
of all would be step function.

A step function is binary in nature, so it will typically output a 1 if the input is higher
than a certain defined threshold, otherwise its output will be 0.

An example would be,

Input 1 ( x j = 0.2
Input 2 (x2) = 1.2
Weight 1 = 0.6
Weight 2 (w2) = 0.4
Threshold = 0.5

Let's calculate the weighted sum now:

x1zv1 + x2w2 = (0.2 x 0.6) + (1.2 x 0.4) = 0.6
In this example, since the total input is higher than the threshold, the neuron fires.

How perceptrons get tra in e d

Remember when you were young and your parents taught you how to identify
a dog on the street. You were told a couple of times that a dog looks like this,
a dog barks like this and pretty soon you could recognize a dog on your own on
the street. Right? When you saw other animals fo r the first time, you could clearly
differentiate between them and a dog.

138 E ssentials of A rtificial Intelligence- I X

Similarly, perceptrons are trained from input training set and the input vectors are
presented to one perceptron after another with weights being modified as below:
For all inputs i,

F(i) = F(i) + l*(T-0 )*N(i), where I is the learning rate

Note: The full equation is F(i) = F(i) + l*g ’(sum of all inputs)*(T-0 )*N(i), where g’
is the derivative of the activation function.
Here, F is the weight vector, N is the input vector, T is the target output that the
perceptron should have known and O is the output provided by the perceptron.
The perceptron is said to have learnt completely when an entire set of inputs is
passed through all training vectors and output is provided without any error.
At this time, if an input vector N (already in the data set used in training) is given
to the perceptron, it will output the correct value as required. If N is not in the
training set, the network will respond with an output similar to other training
vectors close to N.
The perceptron adds all the inputs and separates them into tw o broad
categories— those that cause it to fire (above the threshold) and those that
don't. In this sense, it attempts to draw a line:
n1x1 + n2x2 = t, where t is the threshold. The line divides the points on either side
making tw o distinct categories. As the weights and thresholds can be valued, this
would make a line across the two-dimensional input space. (An example is shown
in the next section). (Source: Medium)

Limitations of the Perceptron Model

Not every set of inputs can be divided by a line like this. Those that can be divided
are called linearly separable (See Fig. 6.5(A)). If the vectors are not linearly separable,
learning will never reach a point where all vectors are classified properly. The most
famous example of the perceptron’s inability to solve problems with linearly non-
separable vectors is the famous Boolean XOR problem.
Bjaj& B To learn more about this problem,
Scan QR code or visit:
HoiSte https://www.youtube.com/watch?v=Fp33mf3L6AE

It is a lecture by Prof. S. Sengupta, Department of Electronics and Electrical

Communication Engineering, NT Kharagpur.

G etting S tarted with Neural Networks 139

 t A 4 B

Fig. 6.6: Linearly separable and non-separable data (Source: Sebastian Raschka)

Neural Networks are trained typically in tw o ways:

REINFORCEMENT LEARNING
In Reinforcement Learning, algorithm learns from its interaction with the
environment. It is a type of dynamic training that trains algorithms using a system
of reward and punishment.

As an example, let us say that during input data training with a given set of inputs
is presented to the Neural Network, the corresponding Output is 0.55, while the
target output was expected to be 1.2 (The output calculation is done as per the
formulae defined in perceptron training model on the previous page.).

In this case, the error (1.2 - 0.55) is used fo r training ("w rong by 0.65” ). This is
called punishing the algorithm. In case there are tw o outputs, then the total error
is summed up to give a single number (typically sum of squared errors).

The algorithm improves with more and more data points to move closer to the
target value. The idea is to minimise the error function.

It is interesting to note that we only need to know how much we are wrong but
not in which direction. It is like if you have to go from Delhi to Mumbai, which is
approximately 1500 km, you are going from North to South. Let us say you went
200 km in the wrong direction and you ended up in Punjab, which is north of Delhi.
The algorithm tells you that you are 200 km wrong or 1700 km away from Mumbai
but doesn’t tell you in which direction.

140 Essentials of A rtificial Intelligence- I X

SUPERVISED LEARNING
Supervised Learning is a more structured way of training the algorithm wherein the
neural network, if given more information and training, will tell you how much right
or wrong you are and in which direction. Taking the above example, it will tell you
that you are in the opposite direction of Delhi to Mumbai route and, hence, wrong by
200 km., thus, 1500 km + 200 km = 1700 km away.

Backpropagation is the essence of neural net training. It is the practice of fine-

tuning the weights of a neural net based on the error rate (i.e., loss) obtained in
the previous epoch (i.e., iteration). Proper tuning of the weights ensures lower
error rates, making the model reliable by increasing its generalization.

In order to train a multi-layer neural network, we follow this approach:

• Define random weights (and biases) in the Neural Network.
• Use one or more data of the training set to see how the output(s) is compared
to what they should be (target output) and calculate the error.
• Alter the weights and try a few combinations to get improvement on outputs.
• Iterate with new training data sets and keep tinkering with the weights each
time.
• Keep iterating till the error or loss function is minimized.

To summarize, Neural Network can be trained by supervised or reinforcement

methods but the key to getting better results is to use Backpropagation as it allows
you to iteratively adjust the weights till the algorithm learns to optimize results.

Compare Human Brain and ANNs

Now, according to your understanding, derive an analogy between the components
of an Artificial Neural Network with the Human Nervous System.

I Artificial Neural Network Human Brain and Nervous System 1

Input Layer
Nodes
Output Layer
j

G etting S tarted with Neural Networks 141

' Using Google Experiments, explore the layers of a neural network with your
camera.

Scan QR code or visit:

||g || https://experiments. with google, com/what-neural-nets-see
This experiment lets you turn on your camera to explore what neural nets see, live,
using your camera. Watch the video explainer above to see how each layer of the
jie u r a l nets works._ _ _ _ _ _ _ _ _ _ _ _ _________________________________________

Draw together with a neural network

This experiment lets you draw together with a recurrent neural network model
called Sketch-RNN. The neuralnet is trained to draw by training it on m illions of
doodles collected from the Quick, Draw! game. Once you start drawing an object,
Sketch-RNN will come up with many possible ways to continue drawing this object
based on where you left off. The model can also mimic your drawings and produce
sim ilar doodles. It is just another example of how you can use machine learning in
fun and creative ways.
j® p |j|j] Scan QR code or visit:
|||jij| httpsV/experiments.withgoogle.com/sketch-rnn-demo

DIFFERENCE BETWEEN NEURAL NETWORKS

AND CONVENTIONAL COMPUTING
A conventional computing device works on a serial processing logic wherein a
central processor addresses memory locations array where data and instructions
are stored. The processor reads an instruction as well as any data the instruction
requires from memory addresses and only then the instruction is executed with
results saved in a specified memory location.

In a serial system (even in a standard parallel one as well) the computational

steps are deterministic, sequential and logical with the state of a given variable
tracked from one operation to another.,

142 E ssentials of A rtificial Intelligence- I X

ANNs, on the other hand, are not sequential or necessarily deterministic but are
more probabilistic. There is no central processor but many small ones whose input
is weighted sum and calculated. ANNs respond in parallel (either simulated or
actual) to the pattern of inputs presented to it and do not execute programmed
instructions. There is no memory address for sorting data or instructions. Information
is maintained in the activation State of the network. (Source: University of Wisconsin— Madison)

Biological neurons or nerve cells Silicon transistors

200 billion neurons, 32 trillion 1 billion bytes RAM, trillions of bytes

interconnections on disk

Neuron size: 1 0 - 6 m Single transistor size: 1 0 - 9 m

Energy consumption: 6 - 1 0 joules per Energy consumption: 1 0 -1 6 joules per

operation per second operation per second

Learning capability Programming capability

Where Neural Networks should be used

Neural networks work on approximations and iterative improvements so they

work well in situations where we have a high tolerance to errors and systems
can enhance with time through pattern recognition.
Some of these situations could be:

• Identification of trends and anomalies in a set of patterns

• In extremely diverse data sets with lots of underlying parameters

• Variables correlations are unclear

• Conventional correlation approaches don't work

G etting S tarted with Neural Networks 143

WLTumvm*
ILet’s
p t ’s sstart
tart with generating new strokes based on your handwriting input

Scan QR code or visit:

https://distill.pub/2016/handwriting/

Untrained neural network models are much like new-born babies: They are ignorant
of the world (if considering tabula rasa epistemological theory), and it is only through
exposure to the world, i.e., a posteriori knowledge, that their ignorance is slowly
revised. Algorithms experience the world through data— by training a neural network
on a relevant data set, we seek to decrease its ignorance. The way we measure
progress is by monitoring the error produced by the network.

Neural Network to Cartoonify your image

g can qr CQCj e Qr vjs jf;
https://experiments. with google, com/cartoonify
if
Turn a photo into a cartoon with a neural network.
Cartoonify turns your photo into a cartoon drawing, powered by Draw This by
Dan Macnish and the Google QuickDraw data set.
For best results, try using images with clear objects on a solid background.
V_____________________________________ _____________________________________

\ Objective Type Questions

A. Fill in the blanks.

1. Conventional Computing i s .......................................in nature while Neural Computing
i s .......................................in nature.
2. A neuron collects inputs from all other neurons using ........................................ and
then the neuron sums all the inputs and if the resulting value is greater than a
................................. , it fires.
3. A loss function defines th e .......................................of the neural network with respect
to a training data set and sample output.
4. Backpropagation calculates th e ...................... and propagates it back to earlier layers.
5.................................. determines how quickly a neural network updates its parameters.
6. Tensorflow is an example o f .......................................frameworks.
7........................................... is a class of deep learning models applied in Computer Vision.
8. Deep Learning requires....................................... amounts of training data.

144 Essentials of A rtificial Intelligence- I X

B. State whether the following statements are True or False.
1. No processing occurs at input layer of an ANN.
2. A multilayer perceptron consists of more than one perceptron.
3. The goal of deep learning model is to minimize the loss function during training.
4. Hidden nodes are responsible for data transfer from network to outside world.
5. Activation function translates output into input.
6. A perceptron is a linear model.
7. High-dimensional data means where lot of inputs and related outputs exist.
8. Artificial Intelligence is a subset of Deep Learning.
9. Dendrites receive signal from other neurons.

C. Multiple Choice Questions (MCQs):

1. What is the capacity of a neural network?
4a) Number of patterns it can store (b) Number of outputs
(c) Number of nodes (d) Number of inputs
2. Which of the following is not an activation function?
(a) Linear (b) Beta (c) Unit (d) Logistic
3. A Perceptron consists of:
(a) Weights (b) Summation processor
(c) Activation function (d) Neuron
4. Is the data in this graph linearly separable?
t k

• •

• •
--------------------►
(a) Yes (b) No
5. The number of nodes in the input layer is 6 and the hidden layer is 10. The maximum
number of connections from the input layer to the hidden layer is:
(a) 16 (b) 60 (c) 30 (d) 0.6
6. Assume a simple neural model with 3 neurons and inputs = 1,2,3. The weights to the
input neurons are 2, 4 and 6 respectively. Assume the activation function is a linear
constant value of 5. What will be the output?
(a) 80 (b) 140 (c) 28 (d) 72
7. Which function decides if the neuron should be activated or not?
(a) Bias Function (b) Trigger Function
(c) Activation Function (d) Start Function

G etting S tarted with Neural Networks 145

 8. Which layer performs computation in a multilayer perceptron?
(a) Input Nodes (b) Hidden Nodes
(c) Auxiliary Nodes (d) Fibre Nodes

: Subjective Type Questions I

1. Explain the benefits of Backpropagation and how it can improve the learning algorithms.
2. What is the key difference between Supervised and Unsupervised Learning?
3. How do you think Artificial Neural Networks vary from Human Brain?
4. Do you think Perceptron Model has any constraints? If yes, explain.
5. Describe the application of neural networks in object classification in a self-driving car.
6. Think of examples where Neural Networks can be applied in you school, fo r example,
in school library. What would be the use case and how would you apply it?
7. What is the significance of loss function in the context of neural network training?
8. What are the different parts of the multilayer perceptron model?

M .
Quick Activity
Watch how neural networks see images:

Scan QR code or visit:

https://www.youtube.com/watch?v=GuOM kmynW kw&feature=embJogo

Play with a real neural network in the browser. Learn more about it as you play with the input variables.

Scan QR code or visit:

https://playground. tensorflow. org

146 E ssentials of A rtificial Intelligence- I X

 o
* </> *

Basics of Python for

use in Al P rojects

Learning Objectives:
[p I Acquire introductory Python programming skills
|p | Understand how Al projects can be run using Python

B ackg ro und a n d C o ntext

“ Python has been an important part of Google since the beginning, and remains so as
the system grows and evolves. Today dozens of Google engineers use Python, and we’re
looking for more people with skills in this language.”
— Peter Norvig, Director of Search Quality at Google, Inc.

You can understand and appreciate the importance of Python as a language and
how much it is in demand globally. If you are well equipped in Python, a host
of opportunities open up for you and employability increases significantly. For
designing and implementing Artificial Intelligence and Machine Learning projects,
Python programming is a critical skill and you must master it.

We will start with the basics of Python language and explore how it can be
implemented in AI/ML projects.

wp# .....................................
Programming Language is a language to write programs. It is a set of
instructions that produces various outputs.

A high-level language (HLL) is a programming language that enables a

programmer to write programs that are independent of a particular type of
computer. Such languages are considered high-level because they are closer
to human languages and further from machine languages.

fl
Basics of Python for use in Al P rojects 147
WHAT IS PYTHON
Python Programming Language is a high-level programming language created by
Guido van Rossum in 1989. It resulted in a great general-purpose language capable
of creating anything from desktop software to web applications and frameworks.

Why was Python created

In his own words, Guido revealed that he started working on it as a weekend
project utilizing his free time during Christmas of December 1989. He originally
wanted to create an interpreter which later turned out to be Python, gradually
evolving into a full-fledged programming language.

How the name Python came about

Guido initially believed that the UNIX/C hackers were the target users of his project.
Also, he was fond of watching the famous comedy series— The Monty Python's
Flying Circus. The name Python captured his mind as not only did it appeal to his
taste but also his target users.

A general-purpose programming language is a programming language

designed to be used for writing software in the widest variety of
application domains. Conversely, a domain-specific programming
language is one designed to be used within a specific application domain.

PYTHON PROGRAMMING DOMAINS

• Web Application Development
Python is majorly used in the field of web development, being the preferred
language of many large projects. Key web application frameworks include
Django, Flask, CherryPy and Bottle, which are commonly used and have a large
developer community.

These frameworks are very handy when it comes to simplifying tasks related
to content management and configuration, accessing a database, and handling
network protocols like HTTP, SMTP, FTP and POP.

148 E ssentials of A rtificial Intelligence- I X

• Data Science and Machine Learning

As you have read in Chapter 1, Data Scientists are in so much demand globally
that 40% vacancies are still lying vacant in AI/ML jobs. Pattern Recognition using
sophisticated algorithms and making sense of data are hot skills in demand
today. Python makes it extremely easy to get started with readymade libraries
and large community support.

It has tools and models for web scraping, data cleaning and standard algorithms.

Why You Should Learn Python Programming

Writing code in Python is so intuitive and easy that you do not feel that it is a
lengthy process. You also do not have to remember complicated syntax. It is also
easy to apply in Al and ML applications as well. Some of the other reasons why
you should learn Python are:

• Non-restrictive Syntax
Python can be easily understood by someone who does not know programming
very well. It is a readable and intuitive language which lets you focus on problem­
solving and does not involve complicated syntax or structures.
Let us consider an example of calculating Simple Interest on a given Amount,
Principal and Rate of Interest (%):

Basics of P ython for use in Al P rojects 149

• No Declaration
You do not need any specifiers for declaration of variables. You can define
without mentioning the data type (integer, etc.).
There is no need to use a separator like a to mark the end of a statement.
Indentation replaces brackets to group a block of code and you can use either
tabs or spaces to indent the code.

• Inbuilt OOP Support

Object-Oriented Programming (OOP) is already built into the Python language.
This is a model which can create objects and define relationships between
them. OOP is a bottom-up problem-solving technique that provides a high-level
solution to a problem in the beginning and implements it later.
Python helps you minimize the size of the code that you write and solve it
quickly. You will see later how much you can achieve with small lines of code
in Python.

• Strong Tools in Debugging

If you are building a product, you know how important it is to do debugging
to ensure quality and reliability of the code.
Python provides inbuilt tools such as pdb, pudb, and PyDebug that make
debugging very easy. These tools can pinpoint the exact problem and guide
you how to solve it.
If you have ever worked with Java or C, you will notice that Python breaks the
ideas and biases you associate with typical programming languages.

Python is a high-level and Object-Oriented Programming language which,

according to Stack Overflow, a question-answer website for professionals
and enthusiast programmers, has over 10 million registered users.
According to the website’s survey, Python’s popularity surpassed that of
C# in 2 0 1 8 and PHP in 2 0 1 7 . On the GitHub platform, Python surpassed
Java as the second-most used programming language.

150 E ssentials of A rtificial Intelligence- I X

 . . . . . . . . . . . . . . . . .

Notable Features of Python:

• Python uses an elegant syntax, making the programs that you write
easier to read and understand.
• Python is an easy-to-use language that makes it simple to get your
program working. It is ideal for prototype development and other ad hoc
programming tasks, without compromising on maintainability.
• Python comes with a large standard library that supports many common
programming tasks such as connecting to web servers, reading and
modifying files, etc.
• Python’s interactive mode makes it easy to test short snippets of code.
There is also a bundled integrated development environment called IDLE.
• It can be easily extended by adding new modules implemented in a
compiled language such as C.
• It can also be embedded into an application to provide a programmable
interface through APIs.
• Python runs anywhere, including Mac OS X, Windows, Linux and Unix,
with unofficial builds also available for Android and iOS.
• It does not cost anything to download or use, or to include it in your
application. Python can also be freely modified and redistributed because
while the language is copyrighted, it is available under an open-source licence.
Some unique programming language features of Python are:
• A variety of basic data types are available: numbers (floating point,
complex and unlimited-length long integers), strings (both ASCII and
Unicode), lists, and dictionaries.
• Python supports OOP with classes and multiple inheritances.
• Code can be grouped into modules and packages.
• The language supports raising and catching exceptions, resulting in
cleaner error handling.
• Data types are strongly and dynamically typed. Mixing incompatible
types (e.g., attempting to add a string and a number) causes an
exception to be raised, so errors are caught sooner.
• Python contains advanced programming features such as generators
and list comprehensions.
• Python’s automatic memory management frees you from having to
manually allocate and free memory in your code.
(So u rce : https://wiki.python.org/)
r,Jf»w W W ss(Bw ^Si

Basics of Python for use in Al P rojects 151

PYTHON VIS-A-VIS OTHER LANGUAGES
Stack Overflow calls Python as the fastest growing programming language.

Fig. 7.1: Python as the fastest growing programming language (Source: stackoverfiow)

Python is easy to learn and code. Python programming language runs on a variety
of platforms like Windows, Linux, Macintosh, Solaris, etc.

Python’s readability makes it a great programming language as no time is wasted

with confusing syntax. For instance, look at the following code to print “ hello
world” in Java and Python.
JAVA

PYTHON

Fig. 7.2: Python code is simpler (Source: PluralSight)

152 E ssentials of A rtificial Intelligence- I X

Python is a very popular language in multiple domains like Automation, Big Data,
Artificial Intelligence, Web Development as well as Data Wrangling, Exploration
and Visualization, etc. It is used by a vast number of companies around the globe,
including Google, Yahoo, Facebook, YouTube, IBM, etc.

Google O yahoo/ O YouTube I1M

These advantages make Python a great choice for both beginners and experts.

PYTHON INTERPRETER
Python uses an interpreter which is a computer program that executes as per
program. It performs instructions as written in a programming or scripting language,
without requiring them to be compiled into a machine language program. So, unlike
Java, Python uses an interpreter.

PROGRAMMING ENVIRONMENT
A programming environment is the computer you work on and all the software
applications that are installed on your system which help you write, debug and
run programs. Some environments are better for programming than others but
the best system to learn on is probably the one you are using right now. Let us
see how we can set up our environment on a system.

OVERVIEW
Our goal is to help you get Python up and running on your computer so that you
can write and run your own programs. To do this, we want to:

• Find out if Python is already installed on your computer.

• Install Python if it has not already been installed.

• Install a text editor that will make it easy to run your first program.

• Help you enter and run Hello World, your first Python program.

• Congratulate you if it works and give you some troubleshooting options if it

doesn’t.

Basics of P ython for use in Al P rojects 153

PYTHON INSTALLATION
Let us now learn how to install Python in Windows OS.

1. Go to www.python.org/downloads/
| *5* Download Python | Python.org X
Co to python.org/downloads/ * — r Q ft PythonSMfrwaw! Te u walM iew jUS) j [h ttp s /w w w -pyth on .org/d ov. nlcads, |
■|| Apps (<$) Computer Science pujamcsdwariai - Cl Q MiNDSTCRMS for T CBSE, Class XI Comp (3

Select the latest Python

version for Windows

Fig. 7.3: Python Download

2. A fter the application file is downloaded, we can install it by opening it. This
will open the Python 3 .7 .0 ( 3 2 -bit) Setup window as shown below. Put check
on “ Add Python 3.7 to PATH” .
Python 3.7.0 (32-bit) Setup O

Install Python 3 .7 .0 (3 2 -bit)

Select Install Now to install Python with default settings, or choose
Customize to enable or disable features.

# Install Now
Click on Install Now** C:\Users\preeti\AppData\locar\Pfcgrams\Python\Python37-32

includes IDLE, pip and documentation

Citatcs shortcuts and file associations

•> Custom ize installation

Choose location and features

python
y : lr s t .ll l.ir r s r h .. < n r a ll . . c a r t f r » r - r s m r r r . Q r i ^ t

© Select ‘Add p',thon • I w indows 3 Add Python 3.7 to PATH Cancel

3.7 to PATH’
Fig. 7.4 : Install Python

154 E ssentials of A rtificial Intelligence- I X

3. When the installation is complete, you will see a screen displayed as under.

Fig. 7.5: Python Setup

4. Click Close to finish.

5. As you will see, the computer interface, IDLE (Python 3.7 32-bit), gets installed
(as shown in the figure below) and is ready to use now. You can now write
Python code in the newly installed Python Shell (Editor) or Script window.

A p p S by name v

Python 3.7

iger (M
rrv m
OpenOffice Draw Ql IDLE (Python 3.7 32-bit) n ew

rz?m
Wizard -j OpenOffice Impress Python 3.7 (32-bit)

ce TBp OpenOffice Math g? Python 3.7 Manuals (32-bit) n ew

Ksm l
— OpenOffice Writer Python 3.7 Module Docs (3 new

Fig. 7.6: Start Python IDLE

Basics of Python for use in Al P rojects 155

 How to Use the Python IDLE Shell

ln:1 Col: 0

Python 3.6.5 Shell - ° E E

File Edit Shell Debug Options Window Help

Python 3.6.5 (v3.6.5:f59c0932b4, Mar 28 2018, 16:07:46) [MSC v . 1 9

0 0 32 b i t ( I n t e l ) ] o n W i n 3 2
Type "copyright", "credits" or " l i c e n s e ()" for more information.
» > "Hello World"
'Hello World'
> »

In : 5 Col: 4

Fig. 7.7: Python IDLE Shell

Every Python installation comes with an Integrated Development and Learning

Environment — IDLE or IDE. Python IDLE is very basic in structure and is considered
ideal for a new programmer.
Python IDLE comes with Python installations on Windows and Mac. If you’re a
Linux user, then you should be able to find and download Python IDLE using your
package manager. Once you’ve installed it, you can then use Python IDLE as an
interactive interpreter or as a file editor.
Next, take a look at the menu bar. You will see a few options for using the shell:
You can restart the shell from this menu. If you select Shell Debug Options
that option, then you will clear the state of the shell. View Last Restart F6

It will act as though you have started a fresh instance Restart Shell ^ F6

of Python IDLE. The shell will forget about everything Interrupt Execution
from its previous state.

PYTHON FUNDAMENTALS
To understand the basics of Python, we will cover:
a. Python Variables b. Python Data Types
c. Python Operators d. Python Conditional Statements
e. Python Loops f. Python Functions

156 E ssentials of A rtificial Intelligence- I X

 A data type is simply a structure of data. It is a data storage form at that
can contain a specific type or range of values. Some commonly used data
types include integers, floating point numbers, characters, strings and
arrays.
• * * * ' * • * • * * • * * * * * * < * * * » * > ? * * * * * * * * * * * * * * * ! * « * • * # ^ ( ! ft ^ * t * * * » M » , * * ? ’ J * ^ f * ? * * , W * « # - * • ’• ' * * * # * * * < * < ( * * * , * * 5 , # * • * * ♦ *

Here are some sample programs to get you started in Python:

Scan QR code or visit:

https://wiki.python.org/moin/SimplePrograms

To run your code and check for improvement ideas, have a look at:

Scan QR code or visit:

https://www.pythonchecker.com

VARIABLES IN PYTHON
Variables reserve memory locations to store values. What this essentially means
is that when you create a variable, you reserve some space in memory. Unlike in
languages like Java, C, etc., in Python, there is no need to declare variables before
using them. The declaration happens automatically when a value is assigned to a
variable. The (=) sign is used to assign values to variables. For example,

S = 15

print (S)

This will assign value ‘15’ to the variable ‘S’ and will print it.

Useful Resource
The Python Package Index (PyPI) is a repository of software for the Python
programming language.

Scan QR code or visit:

https://pypi.org

Basics of P ython for use in Al P rojects 157

Python offers many choices for web development:
• Frameworks such as Django and Pyramid
• Micro-frameworks such as Flask and Bottle
• Advanced content management systems such as Plone and Django CMS

Python’s standard library supports many internet protocols:

• HTML and XML
• JSON
• Email processing
• Support for FTP, IMAP and other internet protocols
• Easy-to-use socket interface

Package Index has still more libraries:

• Requests, a powerful HTTP client library
• BeautifulSoup, an HTML parser that can handle all sorts of oddball HTML
• Feedparser, for parsing RSS/Atom feeds
• Paramiko, implementing the SSH2 protocol
• Twisted Python, a framework for asynchronous network programming

Python is widely used in scientific and numeric computing:

• SciPy is a collection of packages for mathematics, science and engineering.
• Pandas is a data analysis and modelling library.
• I Python is a powerful interactive shell that features easy editing and recording
of a work session, and supports visualizations and parallel computing.
• The Software Carpentry Course teaches basic skills for scientific computing,
running bootcamps and providing open-access teaching materials.

DATA TYPES
In Python, all data values are represented by objects. Each object or value has a
data type. Python supports various data types such as numeric, list, tuple, string,
set, dictionary. These data types define the operations possible on the variables
and their storage method. The different kinds of data types are shared for your
reference in the toolbox.
Let us discuss each data type in detail.

158 E ssentials of A rtificial Intelligence- I X

Numeric

It stores numeric values. They are immutable which means that their value cannot
be changed. Python supports three Numeric data types:

Integer type: It can hold all the integer values, i.e., all the positive and negative
whole numbers, for example, 11.
Float type: It holds real numbers. It is represented by decimal and sometimes
scientific notations with E or e, indicating the power of 10 (3-5e2 = 3.5 x 102 = 357),
for example, 6.24.

Complex type: These are of the form x + aj, where x and y are floats and j represents
the square root of -1 (which is an imaginary number), for example, 10 + 6j.
It is easy to convert the integer value to a float value and vice versa using Python.
Let us look at an example below:
A = 11
# Convert it into float type
B = float(A)
print(B)

This code will convert an integer value to a float type. Following is an example of
converting float to integer type:
A = 11.56
# Convert it into int type
B = int(A)
print(B)

List

Lists in Python are like Arrays in C. But in a Python List, you can store elements of
different types whereas in an Array, all the elements should be of the same type.
List is actually the most versatile data type available in Python. It can be written
as a list of comma-separated values between square brackets.
Consider the example below.
States = ['Punjab', 'Haryana', 'Rajasthan', 28]
print(States)

Notice that the States list contains both words and numbers.

Basics of Python for use in Al P rojects 159

Tuples
A Tuple is defined as a sequence of immutable Python objects. Tuples are sequences
like Lists. However, Tuples cannot be changed unlike Lists. Tuples use parentheses
whereas Lists use square brackets.
Consider the example below:
Agnitio = ('Kartik', 'Salil', 'Himanshu')

You must be wondering what the need of Tuples is when we have Lists. This is
because Tuples are faster than Lists. If a constant set of values is to be defined
which we just want to iterate, then Tuple is used instead of List. Tuple is like List,
but you cannot update, delete or add an element to a Tuple.

Things to Remember
Python is a case-sensitive language. This means that Variable and variable are not
the same. Always name identifiers that make sense and are unique. While c = 5 is
valid, w riting count = 5 would make more sense and it would be easier to figure out
what it does even when you look at your code after a long time. Multiple words can
be separated using an underscore— this_is_a_long_variable.

Strings
Strings are the most popular data types in Python. They are created by simply
enclosing characters in quotes. Python treats single and double quotes equally.
Consider the example given below:
S = 'Welcome to Agnitio'
D = "Agnitio"

Set
A Set is a collection of items without any specific order. Every element in a set
is unique. A set is typically created by placing all the items inside curly braces {},
separated by commas. For example,
Set_l = {4,5,6}

In a Set, every element has to be unique. Try printing the code given below:
Set_2 = {4,5,6,6}
Here, 6 is repeated twice but it will print only once.

160 E ssentials of A rtificial Intelligence- I X

Let us see some set operations:

Union: Union of X and Y is a set of all the elements from both the sets. Union is
performed using | operator. Consider the example given below:
X = {10, 11, 12, 13}
Y = {12, 13, 14, 15}
print(X|Y)

Output:
{10, 11, 12, 13, 14, 15}

Intersection: Intersection of X and Y is a set of elements that are common in both

sets. Intersection is performed using & operator. Consider the example given below:
X = {10, 11, 12, 13}
Y = {12, 13, 14, 15}
print(X & Y)

Output:
{12, 13}

Difference: Difference of X and Y (X - Y) is a set of elements that are only in X

but not in Y. Similarly, Y - X is a set of elements in Y but not in X. Consider the
example given below:
X = {10, 11, 12, 13, 14}
Y = {13, 14, 15, 16, 17}
print(X-Y)

Output:
{ 10, 11, 12}

Symmetric Difference: Symmetric difference of X and Y is a set of elements in both

X and Y except those that are common in both. Symmetric difference is performed
using A operator. Consider the example given below:
X = {11, 12, 13, 14, 15}
Y = {14, 15, 16, 17, 18}
print(XAY)

Output:
{16, 17, 18, 11, 12, 13}

Basics of Python for use in Al P rojects 161

Dictionary
Consider a unique ID given to all citizens of a country. So, for every ID number,
there is a name and few other details attached. Consider this ID number as a unique
‘ Key' and the person's details are the 'Values' attached to that particular Key.

Dictionaries contain these 'Key Value’ pairs enclosed within curly braces and Keys
and Values are separated by Consider this example:

Diet = {'N a m e ' : 'Kartik', 'A g e ' :20 }

Dictionary operations are as follows:

1. Accessing elements from dictionary

Diet = {'N a m e ':'Kartik', 'A g e ':20}
print(Diet['N a m e '])

Output:
Kartik

2 . Changing elements in a dictionary

Diet = {'N a m e ':'Kartik', 'A g e ': 20 }

D iet ['Age'] = 30
D iet ['Address'] = 'Delhi'
print(Diet)

Output:
{'Name': 'Kartik', 'Age': 30, 'Address': 'Delhi'}

OPERATORS IN PYTHON

Operators are the tools which can change the values of the variables. Consider the
expression 7 + 3 = 10; here 7 and 3 are operands and + is an operator.

Python supports the following types of Operators:

Arithmetic Operators
These operators are used to do simple mathematical operations like addition,
subtraction, etc. Assume that A = 100 and B = 200 for the table given on the
next page:

162 E ssentials of A rtificial Intelligence- I X

 Operator Description Example
+ Addition It adds values on either side of the operator A + B = 300
- Subtraction It subtracts the right-hand operator from the
A - B = -100
left-hand operator
M ultiplication It multiplies values on either side of the operator A * B = 20000
/ Division It divides left-hand operand by right-hand operator A / B = 0.5
% Modulus It divides left-hand operand by right-hand operand
B% A=0
and returns the remainder
** Exponent It performs exponential (power) calculation on A * * B = 100 to
operators the power 200
Consider:
a = 10
b = 8
c = 0
c = a +b
print(c) # Python Operators
c = a -b
print(c)
Assignment
c= a *b Operators
print(c)
c = a /b
print(c) mmmm
c = a %b 1dentity
print(c) Operators
a = 2
b = 4
c = a**b
print(c)

Output:
18
2
80
1.25
2
16

Basics of Python for use in Al P rojects 163

Comparison Operators

These operators compare the values on either side of their function and then
decide the possible relation among them.

Assume A = 100 and B = 80.

Operator Description Example

If the values of two operands are equal, then the
(A == B) is not true
condition becomes true.
If values of two operands are not equal, then the
(A != B) is true
condition becomes true.
If the value of left operand is greater than the value of
(A > B) is true
right operand, then the condition becomes true.
If the value of left operand is less than the value of right
(A < B) is not true
operand, then the condition becomes true.
>= If the value of left operand is greater than or equal to the
(A >= B) is true
value of right operand, then the condition becomes true.
<= If the value of left operand is less than or equal to the
(A <= B) is not true
value of right operand, then the condition becomes true.

Consider the example given below:

A = 100
B = 80

if (A == B) :
print ("A is equal to B")
else:
print ("A is not equal to B")

if (A != B) :
print ("A is not equal to B")
else:
print ("A is equal to B")

if (A < B) :
print ("A is less than B")

164 E ssentials of A rtificial Intelligence- I X

 else:
print ("A is not less than B")

if (A > B) :
print ("A is greater than B")
else:
print ("A is not greater than B")

A = 5
B = 20

if (A <= B) :
print ("A is either less than or equal to B")
else :
print ("A is neither less than nor equal to B")

if (A >= B) :
print("A is either greater than or equal to B")
else:
print("A is neither greater than nor equal to B")

Output:
A is not equal to B
A is not equal to B
A is not less than B
A is greater than B
A is either less than or equal to B
A is neither greater than nor equal to B

In the above example, conditional statements (if, else) are used. It means if the
condition is true, then execute the print statement, if not, then execute the print
statement inside else.

Assignment Operators

An Assignment Operator is the operator used to assign a new value to a variable.

B asics of P ython for use in Al P rojects 165

Assume A = 100 and B = 80 for the below table.

Operator Description Example

= Assigns values from right-side operands to C = A + B assigns
left-side operand value of A + B to C
+= Add AND It adds right operand to the left operand and C += A is equivalent
assigns the result to left operand to C = C + A
-= Subtract It subtracts right operand from the left operand
C -= A is equivalent
AND and assigns the result to left operand to C = C - A
*= Multiply It multiplies right operand with the left operand
C *= A is equivalent
AND and assigns the result to left operand to C = C * A
/= Divide AND It divides left operand by right operand andC /= A is equivalent
assigns the result to left operand to C = C / A
%= Modulus It takes modulus using two operands and C %= A is equivalent
AND assigns the result to left operand to C = C % A
**= Exponent It performs exponential (power) calculation on C **= A is equivalent
AND operators and assigns value to the left operand to C = C ** A
Consider the example given below:
A = 100
B = 80
c=0
C =A + B
print(C)
C += A
print(C)
C *= A
print(C)
C /= A
print(C)
C = 20
C %= A
print (C)
C **= A
print(C)

166 E ssentials of A rtificial Intelligence - I X

Output:
180
280
28000
280.0
20
1.26765060022822e+130

Logical Operators

The following are the Logical Operators present in Python:

and True, if both the operands are true X and Y

or True, if either of the operands is true X or Y
not True, if operand is false (complements the operand) not X

Consider the example given below:

X = True
Y -- False
print ('X and Y is ', X and Y)
printf'X or Y is', X or Y)
print ('not X is' , not X)

Output:
X and Y is False
X or Y is True
not X is False

Membership Operators

These operators are used to test whether a variable is found in a sequence (Lists,
Tuples, Sets, Strings, Dictionaries) or not. The following are the types of Membership
Operators:

Operator_____________________Description Example
in True, if value/variable is found in the sequence 5 in x
n o tin True, if value/variable is not found in the sequence 5 not in x

B asics of P ython for use in Al P rojects 167

Consider the example given below:
X = [1, 2, 3, 4]
A = 3
print (A in X)
print (A not in X)

Output:
True
False

Identity Operators
These operators are used to check if two variables are located on the same part
of the memory. Two variables that are equal does not mean that they are
identical.
Following are the Identity Operators in Python:

Operator Description Example

is True, if the operands are identical X is True
is not True, if the operands are not identical X is not True

Consider the example given below:

XI = 'Welcome To Agnitio! '
X2 = 1234
Y1 = 'Welcome To Agnitio! '
Y2 = 1234
print (XI is Y 1 )
print (XI is not Yl)
print (XI is not Y2)
print (XI is X 2 )

Output:
True
False
True
False

168 E ssentials of A rtificial I ntelligence - I X

CONDITIONAL STATEMENTS
Conditional statements are used to execute a statement or a group of statements
when some condition is true.
There are namely three conditional statements— If, Elif, Else.
Consider the flow chart shown below:
START

Statements a

Statements c

Rest of the Code

Let us see how it works:

• First, the control will check the If condition. If it is true, then the program will
execute the statements after the If condition.
• When If condition is not true, then the control will check the Elif condition. If Elif
condition is true, then the program will execute the statements after Elif condition.
• If Elif condition is also false, then the control will execute the Else statements.

X = 10
Y = 12

if X < Y:
print ('X is less than Y ' )
elif X > Y :
print ('X is greater than Y' )
else:
p r in t ('X and Y are equal')

Output:
X is less than Y

B asics of P ython for use in Al P rojects 169

 PYTHON LOOPS
Typically, statements are executed sequentially in a program. The first statement
in a function is executed first, followed by subsequent statements. Sometimes, it
may be required to execute a part of code a number of times. A loop statement
allows us to execute such a statement or a group of statements multiple times as
required. The following diagram illustrates such a loop statement:
START

In the above diagram, first the program will check the condition. If it is true, then
the program will move inside the loop and execute the statements. Next, the
program will again check the condition to see if it is still true, then again it will
execute the statements inside the loop. This process will keep on repeating itself
until the condition becomes false. Once the condition becomes false, the program
control will move out of the loop.
«jf^There are two types of loops:.*- :
• Infinite: When the condition will never become false
• Finite: At one point, the condition will become false and the program control
will move out of the loop
In Python, there are three loops:
• while
• for
• Nested

while Loop
In while loop, first the condition is checked and if it is true, control will move inside
the loop and execute the statements until the condition becomes false. This loop
is used when we are not sure of the number of times the part of code needs to
be executed or when we are unsure about the number of iterations.

170 E ssentials of A rtificial I ntelligence - I X

Consider this example: Output:
count = 0 0
, 1
while (count < 10) :
2
print ( count )
3
count = count + 1
print ("Thank You!") 5

6
7
8
9
Thank You !
for Loop
Just like the while loop, for loop also allows a part of the code to be repeated a
certain number of times. In for loop, the number of iterations is defined, unlike
the while loop.

Consider this example:

veg = ['Brinjal', 'Carrot', 'Onion']
for index in range(len(veg)):
print(veg[index] )

Output:
Brinj al
Carrot
Onion

Nested Loops
A Nested loop basically means a loop within a loop. It can be a for loop inside a
while loop or the other way round. Even a for loop can be inside a for loop or a
while loop inside a while loop. All such possibilities are there.

Consider this example:

count = 1
for i in range (10) :
p r i n t (str (i ) * i )
for j in range (0, i) :
count = count + 1

B asics of P ython for use in Al P rojects 171

Output:
1
22
333
4444
55555
666666
7777777
88888888
999999999

Python Functions
Functions are an easy way to divide your code into useful blocks, making the code
in order, more readable, reusable and efficient. Uses of functions are:
1. Code reusability— any developer can take a code and use it in other applications
2. Code Organization and Documentation
3. Abstraction
4. Extensibility and Scalability

ARTIFICIAL INTELLIGENCE WITH PYTHON

Artificial Intelligence has been around for a long time now. As the demand for Al
is growing, Python is also gaining more and more importance.
As discussed earlier, Python is platform-independent, requires less coding, has
pre-built libraries and is easy to learn. Therefore, Python has become the choice of
language for every core Developer, Data Scientist, Machine Learning Engineer, etc.

Less Code

As you all know, implementing Al involves lots of algorithms. Due to Python’s

support for pre-defined packages, one doesn’t have to code algorithms. You can
check the code as you develop to avoid last-minute testing.

Pre-built Libraries

Python has pre-built libraries to implement various Machine Learning and Deep
Learning algorithms. So, every time you want to run an algorithm on a data set,
you just need to install and load necessary packages with a single command. Some
of these pre-built libraries are NumPy, Keras, Tensorflow, Pytorch, etc.

172 E ssentials of A rtificial Intelligence - I X

Ease of Learning
Python uses a very simple syntax that can be used to implement simple computations
like the addition of two strings to complex processes such as building a Machine
Learning model.

Platform-Independent
Python can run on multiple platforms including Windows, MacOS, Linux, Unix,
and so on. While transferring code from one platform to the other, you can make
use of packages such as Pylnstaller that will take care of any dependency issues.

Massive Community Support

Python has a huge community of users which is always helpful when we encounter
coding errors. Python has multiple communities, groups and forums where
programmers post their errors and help each other.

DEMAND FOR PYTHON IN AI PROJECTS

The main reasons for the popularity of Python in AI projects are:

More Computing Power

AI projects, even with Python, require a lot of computing power since building AI
models involves heavy algorithms, data and use of complex neural networks. With
the invention of GPUs, high-level computations and implementation of complex
algorithms have become possible.

B asics of P ython for use in AI P rojects 173

Data Generation
Over the years, immeasurable amount of data has been generated. Such data
needs to be interpreted, analyzed and processed by using advanced tools.

More Effective Algorithms

In the past decade, we have successfully managed to develop state-of-the-art
algorithms that involve implementation of Deep Neural Networks.

Broad Investment

Since tech giants such as Google, Amazon, Netflix and Facebook have started
investing in Artificial Intelligence, it is gaining popularity leading to a rise in demand
for Al-based systems.

MACHINE LEARNING WITH PYTHON

We can also use Machine Learning with Python.

Problem Statement
Imagine if we were to build a Machine Learning model which will predict whether
AQI (Air Quality Index) will be in the ‘harmful’ category or not the following day.
AQI is a key indicator of air quality and clean air. It has been displaying a lot of
variance in our country as well depending on various factors such as vehicular
pollution, industrial waste, crop burning, industrial pollution, etc.

Target Definition
Let us say we define AQI more than 100 as harmful.

174 E ssentials of A rtificial Intelligence - I X

Data Collection
We need to define data points for all the underlying parameters and their historical data
for the last one year. We need to define the geographical location and time as well.

How to go about it
We need to create a model to predict the target variable, which is “AQIHarmful”.
This target variable (AQIHarmful) will store two values:
1 . Yes: Denotes that AQI will be in the harmful category tomorrow
2. No: Denotes that AQI will not be in the harmful category tomorrow
Therefore, this is clearly a classification problem. The Machine Learning model will
classify the output into two classes, either Yes or No.

Algorithm
To build Classification models in order to predict whether or not AQI will be higher
than 100 based on the underlying model.

Deep Learning with Python

Deep Learning can also be used with Python using existing libraries.

Problem Statement
To study credit card data in a bank to know whether a customer will default on
their outstanding credit card bill for the month.

Target Definition
Let us define the target variable as “Carddefault” which will let us know if the
user will default on his bill payment or not.

Data Collection
We need to collect data set of all the users for the last three years and their
repayment history as well along with CIBIL score and other underlying parameters.

How to go about it
We need to create a model to predict the target variable, which is “Carddefault”.
This target variable (Carddefault) will store two values:
1. Yes: Denotes that the card user will default
2. No: Denotes that the card user will not default

Algorithm
To build a Neural Network that can classify a user as defaulter or not.

B asics of P ython for use in Al P rojects 175

Objective Type Questions j

A. Fill in the blanks.

1 ...........................keyword is used to add an alternative condition to an 'if' statement.
2 ........................................................... a n d ............................. are numeric data types.
3 .............................is used to create a variable 'a' equal to '2'.
4 .............. is used to cast the string variable "a" that is equal to "2" into the integer 2.
5 .............................are reserved memory locations to store values.
6 .............................is a collection of unordered items.
7. In set, every item has to b e ..............................
8...................... , ...................and ....................are three types of loops in Python.

B. State whether the following statements are True or False.

1. The difference between list and tuple is that list is mutable while tuple is not.
2. Lambda function doesn't have statements.
3. Xrange returns the list while range returns the xrange object.
4. If a variable is assigned a new value anywhere within the function's body, it's assumed
as Global.
5. Python doesn't require explicit memory management as the interpreter itself allocates
the memory to new variables.
6. The way of implementing generators is known as iterators.
7. Everything in Python is an object.
8. PEP 8 is a sub-language of Python.
9. Punctuation characters such as @, $, and % within identifiers are allowed.
10. Braces are used in Python to indicate blocks of code.
11. Python variables do not need explicit declaration to reserve memory space.
12. You can assign single value to multiple variables simultaneously in Python.
13. All items belonging to a list can be of different data types.
C. Multiple Choice Questions (MCQs):
1. In Python, what is the one function to output content to the console?
(a) echo (b) output (c) print (d) console.log
2. Which of the following statements is NOT true about Python?
(a) Python can be used for web development.
(b) Python's syntax is much like PHP.
(c) Python can run on any type of platform.
(d) Python can be used to generate dynamic web pages.

176 E ssentials of A rtificial Intelligence - I X

 3. Which of the following is a valid example of List in Python?
(a) sampleList = {1,2,3,4,5} (b) sampleList = (1,2,3,4,5)
(c) sampleList = /1,2,3,4,5/ (d) sampleList = [1,2,3,4,5]
4. What symbol can you use to comment out one line of code?
(a) * (b) (comment) (c) // (d) #
5. How would you call the string variable "a" that is equal to "2" into the integer 2?
(a) castTolnt(a) (b) int(a)
(c) integer(a) (d) castToInteger
6. Which of these is a Python static analysis tool that detects the bugs in Python source
code?
(a) Pytorch (b) PyNum (c) PyChecker (d) PyAI
7. Which of the following is not a mutable built-in type?
(a) Sets (b) Lists (c) Dictionaries (d) Strings
8. What do you use to copy an object in Python?
(a) Copy() (b) Copy.copyO (c) Ctrl+C() (d) CopyList()
9. What is a function inside another function known as?
(a) Nested function (b) Loop function
(c) Infinite function (d) Unloop function
10. What are used to create iterators?
(a) Nests (b) Handlers
(c) Strings (d) Generators
11. What is the assignment of more than one function to an operator called?
(a) Looping (b) Operator Overloading
(c) Nesting (d) Iterations
12. How are errors handled in Python programming?
(a) Sample test (b) Exception Handling
(c) Conditional Blocking (d) Marking
13. Starting an identifier with a single leading underscore indicates that the identifier is:
(a) Public (b) Private (c) Null (d) Local
14. Which of these is not a reserved word?
(a) continue (b) assert (c) try (d) surface
15. Which is the continuation character used to denote that line should continue?
(a) \ (b) I (c) / (d) ;
16. Which symbol is not used to denote strings?
(a) ' (b) ( (c) " (d)
17. Hash sign (#) not inside a string means that it is a
(a) Comment (b) Variable (c) Function (d) Identifier

B asics of P ython for use in Al P rojects 177

 18. Which symbol is used for multiple statements on a single line given that neither
statement starts a new code block?
(a) : (b) ; (c) ? (d) #
19. A group of individual statements which make a single code block is called:
(a) Suite (b) String (c) Function (d) Generator
2 0 . W h ic h n u m e ric a l ty p e is n o t in c lu d e d in P y th o n ?
(a) Int (b) Rational (c) Long (d) Float
21. Tuples are enclosed within:
(a) ( ( ) ) (b) <> (c) { } (d) M
22. Dictionaries are enclosed by which brackets?
(a) {} (b) () (c) [] (d) <>
23. What does float(x) do?
(a) Convert x to an integer (b) Convert x to a floating point number
(c) Convert x to a decimal (d) Convert x to a complex number
24. c -= a is equivalent to:
(a) c=a-c (b) c=c-a (c) c = l-a (d) c = c -a -l
25. *A= operator performs:
(a) Exponential calculation on operators and assigns value to the left operand
(b) Multiplication calculation on operators and assigns value to the left operand
(c) Log calculation on operators and assigns value to the left operand
(d) Weighted mean calculation on operators and assigns value to the left operand
26. What does Logical AND mean?
(a) If either of the operands are true, then the condition becomes true
(b) If both the operands are true, then the condition becomes true
(c) If any of the operands is true, then the condition becomes true
(d) If both the operands are false, then the condition becomes true
27. IN operator means to:
(a) Evaluate to true if variable is not declared
(b) Evaluate to true if variable is not in the list
(c) Evaluate to true if it finds a variable in the specified sequence and false otherwise.
(d) Evaluate to true if variable in the specified sequence is not part of the list
28. Which operators among these have the highest precedence?
(a) a)A (b) +-
(c) is / is not (d) in / not in
29. Which operator means to evaluate to false if the variables on either side of the
operator point to the same object and is true otherwise?
(a) is (b) in (c) not in (d) is not

178 E ssentials of A rtificial Intelligence - I X

 30. #!/usr/bin/python
varl = 'Hello Kartik!'
print "Updated String ", varl[:6] + 'Python'
When you execute the above command, the following result is produced:
(a) Kartik Python (b) Updated Strong - Hello Python
(c) Hello Kartik (d) Hello Python
31. What is the output of 40—40.0?
(a) True (b) False
(c) Not Equal (d) Negative
32. P = 100, Rl = 5, T = 2
I = (*p*R|*T) / 100
Print ("Simple Interest is", I)
What is the correct extension for Python files?
(a) .pt (b) .py (c) .ppt (d) -pyt
33. What is the correct way to create a function in Python?
(a) def myFunction(): (b) createmyFunction():
(c) openmyFunction(): (d) declmyFunction():
34. What is used to remove any whitespace from both the beginning and the end of a
string?
(a) clear() (b) del()
(c) strip() (d) trim()
What is used to replace parts of a string?
(a) rec() (b) replace()
(c) switch() (d) change()
Which of these is a dictionary?
(a) ("tomato", "potato", "onion") (b) {"tomato", "potato", "onion"}
(c) ["tomato", "potato", "onion"] (d) {"name": "tomato", "color": "red’
37. How do you start a while loop in Python?
(a) while x>y: (b) x>y {}
(c) while (x>y) (d) while {x>y}
What is used to stop a loop?
(a) break (b) stop
(c) exit (d) clear
39. We have a set x = {1,2,3,4}, and we execute x.remove(lO). What will happen?
(a) Key error is raised (b) All are removed as 1+2+3+4=10
(c) No exception is raised (d) Method not executed

B asics of P ython for use in At P rojects 179

 40. What is the output of the following code?
eval("2 + 4 * 2")
(a ) 10 (b) 8
(c) 12 (d) 2
41. How do you draw a rectangle centred at 10,10 with width and height as 20, 30
respectively?
(a) Canvas.create.rectangle{10,10,20,30}
(b) Canvas_create_rectangle(10,10,20,30)
(c) Canvas.createrectangle{10,10,20,30}
(d) Canvas.create_rectangle(10,10,20,30)
42. #!/usr/bin/python
str = "this book on artificial intelligence is great";
print "str.capitalize() : ", str.capitalize()
What is the output?
(a) str.capitalize() : This book on artificial intelligence is great
(b) str.capitalize() : this book on artificial intelligence is great
(c) this book on artificial intelligence is great
(d) str.capitalize() : This Book On Artificial Intelligence Is Great
43. def printme( s tr ):
"Prints the passed string into the function"
print str
return;
printmefThis book is great")
(a) This book is great (b) This Book Is Great
(c) Print(This book is great) (d) Print.This book is great

: Subjective Type Questions •

1. Why is Python preferred by web developers?
2. What are the different set operations?
3. Describe various data types in Python with examples.
4. What are the uses of Python functions?
5. How is Tuple different from List?
6. Describe 'while' loop with example.
7. What is dictionary data type? Describe its operations with examples.
8. Why is Python the preferred programme for Al?
9. What are the reasons for the rise in demand for Al?
10. How is Machine Learning implemented with Python?

180 E ssentials of A rtificial Intelligence - I X

 Quick Activity
V M W V

Write Python code to generate the given output:

1
2
3
4
5
6
Good B y e !

Wish to learn more coding in Python? Get started.

Scan QR code or visit:

https://w w w .youtube.com /w atch?v=_uQ rJ0TkZlc& t=9s

There are some libraries available for Python and its use in AI:
1. Tensorflow: This is developed by Google and is very popular among
Machine Learning developers.
2. Scikit-Learn: Scikit-Learn is a Python library associated with NumPy
and SciPy. It is considered good for working on complex problems.
3. NumPy: NumPy is a Python library specifically used for computing
scientific/mathematical data.
4. Theano: Theano is a functional library that effectively calculates and
computes mathematical expressions involving multi-dimensional arrays.
5. Keras: This library simplifies the implementation of Neural Networks.
It also has the best functionalities for computing models, evaluating
data sets, visualizing graphs and much more.
6. NLTK: NLTK or Natural Language TooIKit is an open-source Python
library specifically built for Natural Language Processing, text analysis
and text mining.

B asics of P ython for use in AI P rojects 181

Additional Resources
To gamify the learning of Python, you may use tools such as CodeCombat.

CodeCombat:

Scan QR code or visit:

https://codecom bat. com/play

You will play games here and use commands to execute your moves in the
game.
For example, see below how you are given the option to make moves with your
character in the form of Python commands. Discuss with your friends and play
together:

You can see some examples of common code snippets which can be used in your projects easily:

Scan QR code or visit:

https://skillcrush.com /blog/python-program m ing-exam ples/

182 E ssentials of A rtificial I ntelligence - I X

 EXPLORING THE PYTHON COMMUNITY
As you would have seen around you, the Python community is incredibly rich
and diverse. If you are interested in exploring, here are a few learning resources
and communities you can be part of:

• The Python website https://www.python.org

The main Python website is probably not of much interest to you at this
point but it is a great resource as you start to learn more.

• PyCon https://us.pycon.org/2020/
The Python Conference (PyCon) is an incredible event and the community
welcomes new programmers. These events happen all over the world,
throughout the year. If you can make your way to one of these conferences,
you will learn a great deal and meet some really interesting people.

• PyLadies https://www.pyladies.com
Women and minorities are still under-represented in most technology
fields, and the programming world is no different. That said, the Python
community may well be the most welcoming and supportive programming
community for women and minorities. There are a number of groups
dedicated to bringing women and minorities together around programming
in Python, and there are a number of explicit Codes of Conduct for Python-
related events.

PyLadies is one of the most visible of these organizations. It is a great

resource, so go see what it does and what it has to offer.

• Python User Groups https://wiki.python.org/moin/LocalUserGroups

Wherever there are a number of Python programmers, they will find a
way to get together. Python user groups have regular meetings of Python
users. Go take a look at the list of user groups and see if there is one
near you.

B asics of P ython for use in Al P rojects 183

0} DcodeAI
Create Your Future
DcodeAI - Your Learning Gateway to
AI Projects and Experiments
The best way to learn AI is to try it hands-on and in different scenarios. It may not
always be possible to have access to computation resources needed to run AI models
or get pre-trained models and training data.
This is where DcodeAI can help you. It has a simple and intuitive user interface which
helps you navigate through a personalized Do-It-Yourself AI journey.
You can access the app at:

Scan QR code or visit:

h ttp s :/ / w w w .d c o d e a i.c o m /

You can compete on the leaderboard and collaborate on AI projects with your classmates
and share your AI journeys in class.

(!)

Python
Basics

to P yth o n S trin gs In te g e rs C o m p le x
P ro g ra m m in g N u m b e rs
i.. ilil ii. i $■■1 ]

Community
Projects |

{ Modules

Statistics Python Data Science

Computer Vision Natural Language Processing

184 E ssentials of A rtificial Intelligence - I X

 Al Projects for
Co-Curricular Work
Learning Objectives:
Ip l Integrate the learnings to apply in running end-to-end Al projects
l^p l Design and implement Al projects

B ackgro und and C o n text

“ Systems can have trouble processing large amounts of data but to get Al systems off
the ground, they paradoxically need a lot of data. Data science teams are forced to walk
a tight rope to deliver successful projects using large amounts of data while making
sure the systems can process the specific quantities of information.
To combat these challenges, some 76% of respondents said they sometimes try to
label and annotate training data on their own. More than half (63%) said they even
try building their own labelling and annotation automation technology. Ultimately, 71%
of the teams said they outsource training data and other machine learning project
activities.”
— A recent report by Dimensional Research on behalf of Alegion
By now you know what Al is, how it works, Al project lifecycle and even
programming it using Python. What you need to understand and appreciate
is that implementing live Al projects in real world can be difficult as can be seen
from the issues highlighted above in the research report covering over 200
Al professionals’ experiences.

Al P rojects for C o-C urricular W ork 185

 •JL.

EXPERIMENT 1 • % % f ir !

Inclusive Al project i
T h e re is no d o u b t th a t G o o g le H o m e and G o o g le A s s is ta n t are h e lp in g pe op le w ith d is a b ilitie s
fu lfil ta s k s and b e c o m e m o re and m o re in d e p e n d e n t in th e ir d a ily a c tiv itie s .
B u t to in te ra c t w ith th e A s s is ta n t, y o u need to have g o o d s p e a k in g s k ills . T he A s s is ta n t is
n o t a c c e s s ib le to p e o p le w ith lim ite d o r no s p e a k in g s k ills .
The idea is to create de vice s to let use rs c o n fig u re p h y s ic a l o b je c ts (b u tto n s , RFID/NFC tag s,
o b je c ts de tected by an M L m o d e l, ...) to trig g e r c o m m a n d s to the A s s is ta n t.
W ith th is s o lu tio n , pe o p le w ith d is a b ilitie s w ill be able to in te ra c t w ith th e d e vice s and se rvice s
u s in g th e s ta n d a rd c o n s u m e r p ro d u c t c h a in , a d a p tin g th e “ la s t m ile ” to th e s p e c ia l a b ility
each on e o f us has. T h is m ake s th e s o lu tio n n o t o n ly a c c e s s ib le b u t a lso in c lu s iv e . P eople
w o n ’t be d iffe re n t in te ra c tin g w ith th e A s s is ta n t E c o s y s te m .

Scan QR code or visit:

h ttp s://e x p e rim e n ts. w ith g o o g le , co m / p ro je c t-d iv a

EXPERIMENT 2

Al as Puppetmaster i
A l is n o t o n ly fo r g e ttin g ta s k s d o n e b u t can even be used in th e fie ld o f A rts . Le t us lo o k
a t an e x a m p le here:
C reate s h a d o w p u p p e ts w ith y o u r h a n d s in fro n t o f y o u r la p to p o r p h o n e cam e ra. The p u p p e ts
c o rre s p o n d to one o f th e lu n a r c y c le ’s 12 zo d ia c s y m b o ls and a T e n s o rF lo w m o d e l id e n tifie s
w h ic h a n im a l y o u r h a n d s are fo rm in g . G et it rig h t w ith in th e tim e lim it and y o u ’re g ive n th e
n e x t z o d ia c s y m b o l to try , w ith th e u ltim a te go al o f c o m p le tin g all th e 12 s y m b o ls .

Scan QR code or visit:

h ttp s ://e x p e rim e n ts. w ith g o o g le . c o m / s h a d o w -a rt

E ssentials of A rtificial Intelligence - I X

 A
Talk to Books J
T a lk to B oo ks is an e x p e rim e n ta l new w a y to in te ra c t w ith b o o ks. S peak d ire c tly w ith all o f the
1 0 0 ,0 0 0 b o o k s in o u r in d e x. A s k q u e s tio n s o r m ake s ta te m e n ts . T he A l lo o k s fo r s e n te n c e s
w h ic h are c o n v e rs a tio n a l re s p o n s e s . N ew fe a tu re : F ilte r b y c a te g o ry o f b o o k s !

Sca n QR code or visit:

h ttp s :/ / e x p e rim e n ts .w ith g o o g le .c o m / ta lk -to -b o o k s

As you have seen, there are various applications of Artificial Intelligence. Al impacts
almost every walk of life and can even solve really big global problems. If we
keep advancing at this pace in Al development, how do you think some of these
experiments you have seen would advance or humans and machines would interact
15 years from now?

Al in the Future
Take a fe w e x a m p le s o f h o w A l is im p a c tin g o u r live s to d a y and im a g in e h o w it w ill im p a c t the
sam e s itu a tio n /p ro b le m 15 ye a rs fro m n o w . For exa m ple, to d a y , N e tflix g ive s you pe rso n a lize d
re c o m m e n d a tio n s based on w h a t yo u have w a tc h e d u s in g A l re c o m m e n d a tio n e n g in e s and
even ch a n g e s th e c o v e r o f th e title based on w h a t y o u like to see. H o w th is c o u ld e vo lve in
th e n e x t 15 y e a rs!
H o w w ill e a rly disea se d e te c tio n u s in g A l e vo lve in th e n e x t 15 ye a rs ?
P rep are fiv e use case s in th is m a n n e r.

V
Al P rojects for C o-C urricular W ork 187
 Al in Making Attendance Management Smart at School
E x p lo re h o w y o u can use o p e n -s o u rc e C o m p u te r V is io n lib ra rie s and P y th o n to cre a te an
a tte n d a n c e m a n a g e m e n t s o lu tio n fo r y o u r cla ss. Y ou s h o u ld be able to c a p tu re th e a tte n d a n ce
o f th e s tu d e n ts by th e ir fa ce s w ith o u t th e need fo r a m a n u a l ro ll c a ll. C a lcu la te th e s a v in g s ,
tim e lo s t du e to d is e n g a g e m e n t o f s tu d e n ts an d p ro d u c tiv ity im p ro v e m e n t fo r th e te a c h e r.
Use th e e n tire A l p ro je c t life c y c le a p p ro a c h and c le a rly id e n tify th e s co p e . H o w w ill yo u
c o lle c t data, m o d e l an d e v a lu a tio n p a ra m e te rs ?
A ls o , keep in m in d s itu a tio n s like eve ry y e a r y o u r fa ce e vo lve s, so h o w w ill y o u keep u p d a tin g
th e tra in in g data. N ew s tu d e n ts w ill jo in and s o m e m a y leave. H o w w ill th is im p a c t a tte n d a n c e
m a n a g e m e n t?
Create a Project template

V
You can check various AI/ML project ideas contributed by the community
here. You can also add your projects here to get feedback from the
community. In today’s connected world, it is very important to be part
of like-minded communities and build on each other's work and share
knowledge. Curation of knowledge and building things together for the
greater good of the society should be your focus area.
Scan QR code or visit:
h t t p s :/ / w w w .h a c k s t e r .io / M L / p r o je c t s

188 E ssentials of A rtificial I ntelligence - I X

 A
I %% %

Al Tackles Crime
T a c k lin g c rim e is a b ig p ro b le m th e s e d a ys e v e ry w h e re in
th e w o rld . G o v e rn m e n ts have been p u ttin g in s u rv e illa n c e
s y s te m s an d CCTVs e v e ry w h e re . T h e re has been an
in cre a se in p o lic e p a tro l va n s, SOS and h e lp lin e n u m b e rs .
T hese are all v e ry re a ctive a p p ro a ch e s. Can yo u th in k o f w a ys
in w h ic h A l can help in p ro a c tiv e m o n ito rin g and in te rv e n tio n
to p re d ic t th e lik e lih o o d o f a c rim e and h e lp so lv e it?
Use th e e n tire A l p ro je c t life c y c le a p p ro a c h and c le a rly id e n tify th e s c o p e , h o w y o u w ill
c o lle c t data, m o d e l and e v a lu a tio n p a ra m e te rs .
A ls o , keep in m in d s itu a tio n s like c rim e at n ig h t tim e , c o n s tra in ts o f lin e o f v ie w o f ca m e ra s,
an d w h a t if c rim in a ls are w e a rin g m a sks, etc. A lw a y s th in k o f e x c e p tio n h a n d lin g s c e n a rio s .
L a s t b u t n o t th e le ast, keep in m in d th e b ro a d e c o n o m ic s o f th e p ro je c t as w e ll, o th e rw is e
it m a y n o t be fe a s ib le in th e p ra c tic a l w o rld . A t th is stag e, y o u are e xp e cte d to a p p re c ia te
h o w to d e s ig n and ru n A l p ro je c ts in re a l-life s c e n a rio s .
Create a Project template

V
Al P rojects for C o-C urricular W ork 189
 Al Can Make Money!
A lo t o f p e o p le in v e s t m o n e y in s to c k m a rk e ts to gain
re tu rn s in th e s h o rt te rm as w e ll as lo n g p e rio d s o f tim e
th ro u g h s tru c tu re d m u tu a l fu n d s . A ll th e s to c k m a rk e t data
is ava ilable o n lin e on th e N a tion al S to c k E xchange w e b site .
Even th e h is to ric a l d a ta is in p u b lic d o m a in . Im a g in e th a t
you are a P o rtfo lio M a n a g e r o f a large in v e s tm e n t firm and
have th e re s p o n s ib ility to p ic k s to c k s an d p re d ic t re tu rn s
on in v e s tm e n t. H o w w ill y o u go a b o u t it?
Use th e e n tire A l p ro je c t life c y c le a p p ro a c h and c le a rly
id e n tify th e s c o p e , h o w yo u w ill c o lle c t data, m o d e l and
e v a lu a tio n p a ra m e te rs .
A ls o , keep in m in d s itu a tio n s like d iv id e n d s , s to c k s p lit and fre s h is s u e s o f s h a re s. R esearch
s o m e o f th e s e te c h n ic a l te rm s o n lin e and see h o w th e y im p a c t th e s to c k p ric e s . T h is c o u ld
be a b illio n d o lla r s ta rtu p idea as w e ll!
Create a Project template

V
190 E ssentials of A rtificial Intelligence - I X
 Al Chatbot for E-Governance
T here are th o u s a n d s o f q u e rie s th a t a g o v e rn m e n t d e p a rtm e n t
has to ta ke care of, w h e th e r it is fro m lo ca l o ffic ia ls o r th e
g e n e ra l p u b lic . A s y o u k n o w , c u rre n tly th e a d m in is tra tiv e s ta ff
is re s p o n s ib le fo r h a n d lin g all th o s e q u e rie s and it is n o t th e
m o s t e ffic ie n t w a y as it leads to s lo w re sp o n se tim e . T h in k h o w
an A l-e n a b le d c h a tb o t can be used to h a n d le all th e s e q u e rie s
in te llig e n tly and use b o th te x t and v o ic e -e n a b le d in p u ts .
Use th e e n tire A l p ro je c t life c y c le a p p ro a c h and c le a rly id e n tify th e s co p e , h o w yo u w ill
c o lle c t data, m o d e l and e v a lu a tio n p a ra m e te rs .
A ls o , keep in m in d s itu a tio n s like m u ltip le la n g u a g e s and a c c e n ts o f p e o p le , v o ic e -to -s p e e c h
tra n s la to rs , h o w to p ro g ra m N a tu ra l L a n g u a g e P ro c e s s in g to take care o f m ix e d -la n g u a g e
c o m m u n ic a tio n s , s la n g s , h a lf s e n te n c e s , etc.
A lw a y s th in k h o w A l can be used fo r n a tio n -b u ild in g and fo r th e g re a te r g o o d o f s o c ie ty .
T h is c o u ld be y o u r c o n trib u tio n to e -g o v e rn a n c e .
Create a Project template

V
 Making magic mirror with Tensorflow Machine Learning framework. Watch the introductory video:

Scan QR code or visit:

h ttp s ://v im e o .com /240635807

You can read the code here:

Scan QR code or visit:

https://m edium .eom /@ aallan/a-m agic-m irror-w ith-added-tensorflow -b8fcc5528a6

Smart Recruitment using Al

W h e th e r y o u run a b u s in e s s o r are in G o v e rn m e n t o r A ca d e m ics, y o u w ill end up w o rk in g
w ith people and h irin g th e m as y o u r o rg a n iz a tio n g ro w s . C u rre n tly , h irin g is done m a n u a lly
w ith m u ltip le ro u n d s o f in te rv ie w s (te c h n ic a l/H R ) and lo g is tic s a ro u n d a p p o in tm e n t/in te rv ie w
s c h e d u lin g , etc. T h is leads to u n o p tim iz e d can didate exp erience and h ig h e r c o s t to th e c o m p a n y
to hire a su ita b le ca n d id a te given th a t fo r every 10 0 people in te rv ie w e d , o n ly 1 -2 m ig h t a c tu a lly
be fin a lly selected (as an e xa m p le ).
T h in k h o w A l can he lp in a u to m a tin g th is e n tire p ro c e s s fro m c h a tb o ts fo r lo g is itic s p la n n in g
and s c h e d u lin g to v id e o a n a ly tic s c o m b in e d w ith N LP fo r u n d e rs ta n d in g c a n d id a te re s p o n s e s
and in te rp re tin g th e m .
Use th e e n tire A l p ro je c t life c y c le a p p ro a c h and c le a rly id e n tify th e s c o p e , h o w y o u w ill
c o lle c t data, m o d e l and e v a lu a tio n p a ra m e te rs . A ls o , keep in m in d s itu a tio n s like m u ltip le
la n g u a g e s , h a lf s e n te n c e s , b o d y la n g u a g e bias, re g io n a l bias, etc.
Create a Project template

v
192 E ssentials of A rtificial Intelligence - I X
 A s m o re and m o re ne w te c h n o lo g ie s s u ch
as A l and lo T c o m e in to play, ris k s w ill ge t
m o re c o n c e n tra te d in to a c o m m o n n e tw o rk .
C y b e rs e c u rity b e c o m e s e x tre m e ly c o m p lic a te d
in s u c h s c e n a rio s an d go e s b e yo n d w h ite lis ts ,
b la c k lis ts and fire w a lls an d m o re to w a rd s
d e te c tin g u n u s u a l a c tiv ity and p a tte rn s in c lu d in g
th e m o v e m e n t o f d a ta p a cke ts. T h in k h o w A l
a lg o rith m s can sc ra p e th ro u g h v a s t a m o u n ts o f
lo g s to id e n tify s u s c e p tib le u s e r b e h a v io u r.
Use th e e n tire A l p ro je c t life c y c le a p p ro a c h and
c le a rly id e n tify th e s c o p e , h o w y o u w ill c o lle c t
data, m o d e l and e v a lu a tio n p a ra m e te rs .
A ls o , keep in m in d s itu a tio n s like U se r and E n tity B e h a v io u ra l A n a ly tic s (U E B A ) w h ic h re lie s
on deep le a rn in g m e th o d s and b lo c k c h a in m e th o d s .
Create a Project template

Al P rojects for C o-C urricular W ork 193

 Al as a Judge
In India, w e have 19 ju d g e s fo r every 10 lakh people w h ic h m akes it v e ry d iffic u lt to cle ar
a lm o s t 30 m illio n p e n d in g cases. W e c a n n o t p ro d u c e e n o u g h ju d g e s in s u ch a s h o rt tim e b u t
n e ith e r can ju s tic e be delayed. T h in k h o w it can be m ade p o s s ib le to in te rp re t all p a st cases,
p re ce d e n ts and ju d g e m e n ts by an A l ju d g e w h ic h w ill c o n c u rre n tly solve all cases. In the US,
so m e la w c o m p a n ie s are u sin g A l to p re d ic t the lik e lih o o d o f w in n in g a case.
Use th e e n tire A l p ro je c t life c y c le a p p ro a c h and c le a rly id e n tify th e s co p e , h o w y o u w ill
c o lle c t data, m o d e l and e v a lu a tio n p a ra m e te rs .
A ls o , keep in m in d s itu a tio n s like lo ca l la n g u a g e s in ju d g e m e n t, o ld n o n -d ig itiz e d tra n s c rip ts ,
u n s tru c tu re d d a ta and biases.
Create a Project template

V
Google AIY Voice Kit
The AIY Voice Kit from Google lets you build your own natural language
processor and connect it to the Google Assistant or Cloud Speech-to-Text
service, thus allowing you to ask questions and issue voice commands to
your programs. All this fits into a handy little cardboard cube, powered by
a Raspberry Pi.
Everything you need is provided in the kit, including the Raspberry Pi.
Scan QR code or visit:
https://aiyprojects.withgoogle.eom/voice#project-overview

194 E ssentials of A rtificial Intelligence - I X

 ■ \

Customer Insights using Al

G one are th e days o f c u s to m e r s u rv e y s and fe e d b a c k
fo rm s . N o w , c u s to m e rs e x p re s s th e ir o p in io n a b o u t
a p ro d u c t/s e rv ic e and b ra n d p e rc e p tio n on s o c ia l
m e d ia p la tfo rm s in th e fo rm o f lik e s , c o m m e n ts ,
s h a rin g p ic tu re s and v id e o s a m o n g o th e r fo rm a ts . A ll
th is g e n e ra te s m a s s iv e u n s tru c tu re d d a ta fo r a n yo n e
to m ake sense of. A na lyze h o w A l can in te rp re t th is
data a lo n g w ith s e n tim e n t a n a ly s is to c o m e up w ith
c u s to m e r in s ig h ts . T h in k h o w N LP can be a p p lie d here
to all fo rm a ts o f data.
Use th e e n tire A l p ro je c t life c y c le a p p ro a c h and c le a rly
id e n tify th e sco p e , h o w y o u w ill c o lle c t data, m o d e l
and e v a lu a tio n p a ra m e te rs .
A ls o , keep in m in d s itu a tio n s like te x t in a p ic tu re ,
m u ltip le ty p e s o f c o n te n t e m b e d d e d in to a v id e o , th re a d e d c o m m e n ts , etc.

Create a Project template

V
Al P rojects for C o-C urricular W ork 195
 Fake News Detection using Al
Fake n e w s is b e c o m in g a big m en ace . P eople are u s in g s o c ia l
m e d ia p la tfo rm s s u c h as F aceb oo k, T w itte r, W h a ts A p p , etc., to
spread ru m o u rs and fake ne w s w h ic h leads to w ro n g p e rc e p tio n s
and u n in te n d e d c o n s e q u e n c e s in s o c ie ty . T h in k a b o u t h o w A l
can be used to d e te c t w h e th e r th e n e w s is fake o r real.
Use th e e n tire A l p ro je c t life c y c le a p p ro a c h and c le a rly id e n tify th e sco p e , h o w y o u w ill c o lle c t
data, m o d e l and e v a lu a tio n p a ra m e te rs . O ne can a ls o cre a te d e e p fa ke s n o w and im p e rs o n a te
a n y c e le b rity to v a lid a te a n e w s ite m .
A ls o , keep in m in d s itu a tio n s like re g io n a l la n g u a g e s, v o ic e and te x t ne w s, co d e d la n g u a g e
w ith e m b e d d e d h y p e rlin k s . T h in k h o w p a tte rn re c o g n itio n and N LP can be used here.
A lw a y s th in k h o w A l can be used fo r n a tio n -b u ild in g and fo r th e g re a te r g o o d o f s o c ie ty .
T h is c o u ld be y o u r c o n trib u tio n to e -g o v e rn a n c e .

Create a Project template

V
We have now understood how Al knowledge can be applied in real-life projects and,
most importantly, the fundamentals and approach behind project design. We also
took note of exception handling, having a business case and constraints. We must
always look for issues of ethics and bias in whatever Al project we undertake. Last
but not the least, Al is a tool which will empower you and lead towards not only
significant personal growth and career advancement but also help you contribute
to nation-building and e-governance.

196 E ssentials of A rtificial Intelligence - I X