UNIT 1 – Introduction to

Artificial Intelligence
COCSE58
Harshita Sharma
Lecture Overview

 Intelligence
 Artificial Intelligence
 Characteristics
 Approaches
 Types
 A comparison of natural and artificial intelligence
 Scope and goals
 History of development
 AI today – Applications
 Challenges
 ML, DS, DL
 Task domains of AI
 Knowledge and AI
 Properties of Knowledge
 Knowledge Representation
 Symbolic Processing
 Search
 Heuristic
 Myths
 Future directions/current trends
 Assignment
Intelligence

 Cognitive ability found in living beings (that enable individuals to perceive,

process, understand, and use information).
 It encompasses a broad range of mental faculties and skills that enable
individuals to
 Learn
 Reason
 problem-solve
 adapt to new situations
 interact effectively with their environment
 Also includes aspects of memory and creativity
 Latin word – intelligo – I understand
 Basic meaning of intelligence – the ability to understand
 Some key tasks that are performed using intelligence:
 Plan and define goals
 React dynamically, quickly
 Give appropriate responses
 Handle ambiguity, contradictions, incomplete and uncertain information
 Recognize relative importance of different elements of a situation
 Problem solving
 Learning
 analyzing
Handling contradictions:

 All students passed the exam.

 Ritvik failed the exam.
 Apparently contradictory.
 Intelligent person would deduce: all students except Ritvik passed the
exam
Relative importance of elements

 Ritvik with a red shirt met with an accident.

 Here two things are given in relation to Ritvik
 Red shirt
 Accident
 An intelligent person would give more importance to the element of
accident as compared to the element red shirt.
Artificial Intelligence

 Development to some extent the similar kinds of abilities in a machine.

 The science and engineering of building intelligent machines.
 AI is a branch of computer science that aims to create intelligent machines
capable of performing tasks that typically require human intelligence.
 These tasks include learning, reasoning, problem-solving, perception,
understanding natural language, and more.
 AI currently encompasses a huge variety of subfields, ranging from the
general (learning and perception) to the specific, such as playing chess,
proving mathematical theorems, writing poetry, driving a car on a
crowded street, and diagnosing diseases.
Definition by John McCarthy

 “Artificial Intelligence is the science and engineering to make intelligent

machines, especially intelligent computer programs. It is related to the
similar task of using computers to understand human intelligence, but AI
does not have to confine itself to methods that are biologically
observable."
 This definition highlights the focus of AI on creating machines and
computer programs that can perform tasks that typically require human
intelligence, such as reasoning, learning, problem-solving, and decision-
making. It also emphasizes that AI is not restricted to emulating human
intelligence exactly; rather, it aims to develop intelligent systems that can
solve problems and achieve tasks efficiently and effectively.
Characteristics of AI programs

 Adaptability: AI systems can adapt and improve their performance based

on experience and exposure to new data. They can learn from past
interactions and adjust their behavior to optimize results.
 Autonomy: AI can function autonomously without constant human
intervention. Once trained and deployed, AI systems can carry out tasks
independently, making decisions and taking actions as needed.
 Learning: AI can acquire knowledge and improve its performance through
learning algorithms. Machine learning and deep learning techniques allow
AI systems to identify patterns in data, make predictions, and refine their
understanding over time.
 Reasoning and Problem-Solving: AI can analyze complex information, draw
logical inferences, and solve intricate problems. It can process vast
amounts of data and consider multiple variables to arrive at decisions.
 Natural Language Processing (NLP): AI systems can understand, interpret,
and respond to human language. This ability enables AI to interact with
users through speech and text interfaces.
 Perception: AI can perceive and interpret its environment through sensors
and cameras, allowing it to process visual and auditory information.
 Creativity: Some AI systems demonstrate creative capabilities, such as
generating artistic works, composing music, or designing novel solutions to
problems.
 Planning and Decision Making: AI can create plans and strategies to achieve
specific goals, evaluating potential outcomes and choosing the most
appropriate course of action.
 Pattern Recognition: AI excels at recognizing patterns in data, whether in
images, speech, or numerical data. This skill is vital for tasks like image
recognition and speech recognition.
 Parallel Processing: AI systems can process vast amounts of data and perform
multiple tasks simultaneously, making them highly efficient for complex
computations.
 Emotion Recognition: Some AI systems are capable of recognizing human
emotions through facial expressions or vocal cues, enhancing their ability to
interact with users.
 Big Data Analysis: AI can handle and analyze large datasets, extracting
valuable insights and patterns that would be challenging for humans to process
manually.
 Automation: AI programs can automate repetitive tasks, reducing the need
for human intervention and increasing efficiency.
 Learning from User Interaction: AI programs can learn from user interactions
and feedback, adapting their responses to provide more personalized and
relevant information.
 Robustness and Error Handling: Advanced AI programs are designed to
handle errors and edge cases gracefully, maintaining stability and avoiding
catastrophic failures.
And more depending on the specific application.
Approaches to AI
 we see eight definitions of AI, laid out along two dimensions. The definitions
on top are concerned with thought processes and reasoning, whereas the
ones on the bottom address behavior.
 The definitions on the left measure success in terms of fidelity to human
performance, whereas the ones on the right measure against an ideal
performance measure, called rationality. A system is rational if it does the
“right thing,” given what it knows.
Acting humanly: The Turing Test
approach
 The Turing Test, proposed by Alan Turing (1950), was designed to provide a satisfactory
operational definition of intelligence. A computer passes the test if a human interrogator,
after posing some written questions, cannot tell whether the written responses come from
a person or from a computer.
 The computer would need to possess the following capabilities:
 natural language processing to enable it to communicate successfully in English;
 knowledge representation to store what it knows or hears;
 automated reasoning to use the stored information to answer questions and to draw new
conclusions;
 machine learning to adapt to new circumstances and to detect and extrapolate
patterns.
 Turing’s test deliberately avoided direct physical interaction between the interrogator and
the computer, because physical simulation of a person is unnecessary for intelligence.
Some weaknesses

 Turing deserves credit for designing a test that remains relevant 60 years
later. Yet AI researchers have devoted little effort to passing the Turing Test,
believing that it is more important to study the underlying principles of
intelligence than to duplicate an exemplar.
 The quest for “artificial flight” succeeded when the Wright brothers and
others stopped imitating birds and started using wind tunnels and learning
about aerodynamics. Aeronautical engineering texts do not define the
goal of their field as making “machines that fly so exactly like pigeons that
they can fool even other pigeons.”
 Not all human behaviour is intelligent.
Example: anger
 Some intelligent behaviour might not be possible for humans to exhibit.
Example: some human unsolvable probles are solvable by machines.
Thinking humanly: The cognitive
modeling approach
 Determining how humans think.
 Getting inside the actual working of the human mind.
 There are three ways to do this:
 through introspection—trying to catch our own thoughts as they go by;
 through psychological experiments—observing a person in action; and
 Through brain imaging—observing the brain in action.
The General Problem Solver

 For example, Allen Newell and Herbert Simon, who developed GPS, the
“General Problem Solver” (Newell and Simon, 1961), were not content
merely to have their program solve problems correctly. They were more
concerned with comparing the trace of its reasoning steps to traces of
human subjects solving the same problems.
 The interdisciplinary field of cognitive science brings together computer
models from AI and experimental techniques from psychology to construct
precise and testable theories of the human mind.
Thinking rationally: The “laws of
thought” approach
 Right thinking, the irrefutable reasoning process –Aristotle’s syllogisms (a
form of arguing in which two statements are used to prove that a third
statement is true).
 His syllogisms provided patterns for argument structures that always yielded
correct conclusions when given correct premises—for example, “Socrates is
a man; all men are mortal; therefore, Socrates is mortal.” These laws of
thought were supposed to govern the operation of the mind; their study
initiated the field called logic.
 Emphasis is on correct inferences.
Pittfalls

 There are two main obstacles to this approach. First, it is not easy to take
informal knowledge and state it in the formal terms required by logical
notation, particularly when the knowledge is less than 100% certain.
 Second, there is a big difference between solving a problem “in principle”
and solving it in practice. Even problems with just a few hundred facts can
exhaust the computational resources of any computer unless it has some
guidance as to which reasoning steps to try first.
Acting rationally: The rational agent
approach
 An agent is just something that acts.
 A rational agent is one that acts so as to achieve the best outcome or,
when there is uncertainty, the best expected outcome.
 Making correct inferences is sometimes part of being a rational agent,
because one way to act rationally is to reason logically to the conclusion
that a given action will achieve one’s goals and then to act on that
conclusion. On the other hand, correct inference is not all of rationality; in
some situations, there is no provably correct thing to do, but something
must still be done.
Comparing with other approaches

 There are also ways of acting rationally that cannot be said to involve
inference. For example, recoiling from a hot stove is a reflex action that is
usually more successful than a slower action taken after careful
deliberation.
 All the skills needed for the Turing Test also allow an agent to act rationally.
Knowledge representation and reasoning enable agents to reach good
decisions. We need to be able to generate comprehensible sentences in
natural language to get by in a complex society. We need learning not
only for erudition, but also because it improves our ability to generate
effective behavior.
 The rational-agent approach has two advantages over the other
approaches. First, it is more general than the “laws of thought” approach
because correct inference is just one of several possible mechanisms for
achieving rationality.
 Second, it is more amenable to scientific development than are
approaches based on human behavior or human thought. The standard of
rationality is mathematically well defined and completely general, and can
be “unpacked” to generate agent designs that provably achieve it.
Human behavior, on the other hand, is well adapted for one specific
environment and is defined by, well, the sum total of all the things that
humans do.
Types of Artificial Intelligence

AI (Based on
Capabilities)

Narrow / Weak General/Strong

Super AI
AI AI
 Narrow or Weak AI: Also known as narrow AI or weak AI, this type of AI is
designed to perform specific tasks or solve particular problems. These AI
systems are specialized and excel in a limited domain but lack the ability to
generalize their knowledge to other areas. Examples include virtual
assistants like Siri and chatbots.
 General or Strong AI: General AI, also referred to as strong AI, aims to
exhibit human-like intelligence and the ability to understand, learn, and
perform any intellectual task that a human can do. This level of AI has not
yet been achieved, and researchers are working towards developing
machines with human-level cognitive abilities.
 Artificial Super Intelligence: ASI will be human intelligence in all aspects
from creativity, to general wisdom, to problem-solving. Machines are going
to be capable of exhibiting intelligence that we have a tendency to
haven’t seen within the brightest amongst. This is the kind of AI that a lot of
individuals square measure upset concerning, and also the form of AI that
individuals like Elon musk assume can cause the extinction of the human
race.
Comparing Natural and Artificial
Intelligence
 Similarities
 Problem Solving ability
 Pattern recognition
 Learning and Adaptation
 Task Execution
 Decision making
 Differences
 Flexibility and creativity
 Limitations
 Emotions
 AI: AI systems process data at incredible speeds, handle vast amounts of
information, and excel at specific tasks. However, they lack human-like
consciousness and self-awareness.
 Human Intelligence: Humans possess general intelligence, adaptability,
creativity, and emotional understanding. Yet, they may be limited by
cognitive biases and information processing speed.
 AI is highly precise and can crunch complex data, but it lacks intuition and
context sensitivity.
 Humans have creativity and empathy but may make errors and be limited
in terms of data processing.
Scope and Goals of AI

 The scope of Artificial Intelligence (AI) is vast and continues to expand

rapidly as technology advances. AI has the potential to revolutionize
various industries and domains, transforming the way we live and work.
 AI seeks to build machines that can mimic human intelligence, either by
replicating specific cognitive functions or by achieving general
intelligence.
 The ultimate goal is to develop AI systems that exhibit human-like
intelligence and adaptability.
History

1. Early Concepts (Pre-20th Century):

 The idea of creating artificial beings with human-like characteristics can be
found in ancient myths and folklore from various cultures. Tales of
automatons, golems, and other humanoid creatures captivated the
human imagination.
 In the 13th century, philosopher and scientist Albertus Magnus reportedly
created a humanoid automaton that could perform certain actions.
2. Emergence of Mathematical Logic (17th - 19th Century):
 The groundwork for AI started with the emergence of mathematical logic
and formal reasoning systems.
 Mathematicians like Gottfried Wilhelm Leibniz and George Boole made
significant contributions to symbolic logic, which became essential for AI
research.
3. Early Computational Machinery (Late 19th - Early 20th Century):
 In the late 19th and early 20th centuries, inventors and thinkers created
mechanical devices that imitated human or animal behavior.
 Charles Babbage's analytical engine, designed in the 1830s, laid the
foundation for modern computing and programmable machines.
4. Dartmouth Conference (1956):
 In the summer of 1956, a group of researchers, including John McCarthy,
Marvin Minsky, Nathaniel Rochester, and Claude Shannon, convened at
Dartmouth College for a workshop on AI.
 The term "Artificial Intelligence" was coined during this conference, marking
the formal establishment of AI as a field of study.
 The attendees were optimistic about creating machines that could mimic
human intelligence through symbol manipulation and logical reasoning.
5. Early AI Research (1950s - 1960s):
 In the late 1950s and 1960s, AI research saw significant progress.
Researchers explored rule-based systems and symbolic reasoning to create
intelligent programs.
 Allen Newell and Herbert A. Simon developed the Logic Theorist, a
computer program capable of proving mathematical theorems.
 John McCarthy developed the programming language LISP, which
became a fundamental tool for AI research.
6. AI Winter (1970s - 1980s):
 In the 1970s and 1980s, AI research faced significant challenges and
criticisms due to overinflated expectations and unfulfilled promises.
 Funding for AI research decreased, and progress stagnated, leading to
what is known as the "AI winter," a period of reduced interest and support
for AI research.
7. Expert Systems (1980s):
 During the AI winter, expert systems emerged as a dominant area of
research. These systems utilized rule-based knowledge to simulate human
expertise in specific domains.
 Expert systems demonstrated practical applications in fields like medicine,
finance, and engineering.
8. Emergence of Machine Learning (1990s - 2000s):
 In the 1990s, AI research experienced a resurgence, fueled by
advancements in machine learning techniques.
 Machine learning algorithms such as Support Vector Machines (SVM) and
neural networks gained popularity, enabling AI systems to learn from data
and improve their performance.
9. Deep Learning Revolution (2010s - Present):
 The 2010s marked a significant breakthrough in AI with the widespread
adoption of deep learning techniques.
 Deep learning, a subset of machine learning, utilizes artificial neural
networks with multiple layers to process complex data and make more
accurate predictions.
 Deep learning has revolutionized areas like image recognition, natural
language processing, speech recognition, and autonomous vehicles.
10. AI Applications and Integration (Present):
 AI applications have become integrated into various industries and aspects
of daily life.
 AI-powered technologies, including virtual assistants, recommendation
systems, language translation, robotics, autonomous vehicles, and medical
diagnosis, have become common
AI today : Application Areas

 Automation and Efficiency: AI enables automation of tasks that were

previously time-consuming and labor-intensive. It streamlines processes,
increasing efficiency, and productivity in industries like manufacturing,
logistics, and customer support.
 Data Analysis and Insights: AI's ability to process vast amounts of data
quickly and accurately allows businesses and researchers to extract
valuable insights. This aids in data-driven decision-making and helps identify
patterns and trends that might otherwise go unnoticed.
 Personalization and Customer Experience: AI-powered recommendation
systems and personal assistants enhance customer experiences by
providing personalized content, product suggestions, and support. This
leads to higher customer satisfaction and loyalty.
 Healthcare Advancements: AI is revolutionizing healthcare by aiding in
medical imaging analysis, drug discovery, personalized treatment plans,
and disease prediction. This technology can potentially improve patient
outcomes and reduce medical errors.
 Education and Learning: AI applications in education facilitate personalized
learning experiences, adaptive testing, and intelligent tutoring systems.
These technologies cater to individual student needs, leading to improved
learning outcomes.
 Natural Language Processing (NLP): AI's NLP capabilities have transformed
human-computer interactions. Voice assistants, chatbots, and language
translation services have become integral parts of our daily lives.
 Autonomous Vehicles: AI is a key component in developing self-driving
cars, which have the potential to revolutionize transportation, reduce
accidents, and enhance mobility for various demographics.
 Finance and Fraud Detection: AI algorithms assist financial institutions in
fraud detection, risk assessment, and algorithmic trading, making financial
processes more secure and efficient.
 Environmental Sustainability: AI is utilized in various ways to address
environmental challenges, such as optimizing energy consumption,
managing waste, and predicting natural disasters.
 Scientific Research: AI contributes to scientific discoveries by analyzing
complex data sets, simulating experiments, and assisting researchers in
data-driven hypothesis testing.
 Art and Creativity: AI-driven generative models are now capable of
creating art, music, and literature, blurring the lines between human and
machine creativity.
 Security and Cybersecurity: AI helps in threat detection and cybersecurity
by identifying unusual patterns and behaviors, protecting individuals and
organizations from cyber-attacks.
Challenges

 Ethical considerations
 Bias
 Explainability
 Accountability and autonomy
 Privacy and data protection
 Job displacements
 Adoption and integration
 Interdisciplinary collaboration
 General AI
Some often confused terms
Task Domains of AI
Knowledge and AI

 The relationship between knowledge and AI is fundamental, as knowledge

forms the basis for building intelligent AI systems. AI relies on knowledge
representation, acquisition, and utilization to make informed decisions and
perform various tasks.
 AI systems need to acquire, represent and reason through the knowledge
built into it to make its best use.
Properties of Knowledge

 Indispensable
 Subjective
 Dynamic
 Voluminous
 Hard to characterize accurately
 Organization affects efficiency and usage
 Contextual
 Hierarchical
 Combinatorial
 Uncertainity
Knowledge Representation

 Knowledge representation is a crucial aspect of Artificial Intelligence (AI)

that involves capturing and organizing information in a structured form so
that AI systems can understand, reason, and make informed decisions.
 The goal of knowledge representation is to create a formal and machine-
readable representation of knowledge that can be processed and utilized
by AI algorithms.
 Logic, Semantic networks, frames, ontologies, production rules, etc.
Impact of a good KR scheme

 Facilitates Reasoning and Decision-Making

 Handling Complex Domains
 Knowledge Reusability
 Knowledge Sharing and Collaboration
 Supporting Learning and Adaptation
 Handling Uncertainty
 Human-AI Interaction
 Adaptation
Symbolic Processing

 Symbolic processing in AI refers to a traditional approach that represents and

manipulates knowledge using symbols, rules, and logic. It is also known as
symbolic AI or classical AI.
 In symbolic processing, information is encoded in a structured and explicit form,
allowing AI systems to reason, infer, and make decisions based on logical rules
and relationships.
 While symbolic processing has shown success in rule-based expert systems and
certain AI applications, it also has limitations. It may struggle to handle large-
scale and uncertain data, which are better suited for approaches like machine
learning. Hybrid systems that combine symbolic processing with statistical
learning techniques have emerged to address these challenges, creating a
more comprehensive AI paradigm.
 Symbolic processing in AI emphasizes explicit knowledge representation, logical
reasoning, and rule-based systems to enable AI systems to perform intelligent
tasks, reason logically, and emulate human expertise in specific domains.
Search in AI

 In AI, the search process refers to the systematic exploration of a problem

space to find a solution that meets certain criteria or satisfies a specific
goal. Search algorithms are fundamental components of AI, and they play
a crucial role in problem-solving, decision-making, planning, and
optimization tasks. The search process involves traversing a search space,
which represents all possible states or configurations of a problem, and
identifying a sequence of actions or steps to reach a desirable outcome.
 Search in AI is of significant importance as it forms the backbone of various
problem-solving and decision-making tasks. The ability to search and
explore a problem space efficiently is fundamental to many AI
applications.
A good search technique enhances:
 Problem solving
 Optimization
 Planning and decision making
Crucial to applications like:
 Game playing
 Path finding and routing
 Resource allocation and scheduling
 Constraint Satisfaction problems
 Automated theorem proving
Heuristic

 A heuristic is a problem-solving technique or a rule of thumb that provides a

practical and efficient approach to finding solutions in complex or
uncertain situations.
 In the context of Artificial Intelligence (AI) and search algorithms, heuristics
are used to estimate the cost or potential of states in a search space and
guide the search process towards more promising paths.
 Informed estimation
Myths regarding AI

 Human-like
 Complete replacement for AI
 Completely autonomous
 Infallible
 Singular Entity
 Only for Tech Companies
 Threat to Humanity
 All AI is deep learning
Latest Trends

 Edge computing
 Responsible AI
 Healthcare
 NLP
 Autonomous Vehicles
 Retail and e-commerce
 Education
 Generative AI
 Sustainability
 Cybersecurity
Assignment- 1

 Perform a SWOT Analysis of Artificial Intelligence

 AI in everyday life. Choose any 3 day-to-day applications of AI you use –
identify the specific underlying AI techniques being used, positive impacts,
challenges, ethical concerns, future scope
 Elaborate and compare the terms : AI, ML, DS, DL. How are these related?
 Give counter arguments for any 3 myths regarding AI.
Instructions:
To be submitted in hard copy in presentable form.
Assignment should be elaborate.