ARTIFICIAL INTELLIGENCE

BANGALORE UNIVERSITY – MARCH 2023 (SOLVED QP)

SECTION - A

Answer any 5 questions. Each carries 2 marks. (5 x 2 = 10)

1) List the areas of application of AI.

Ans. 1. Robotics, 2. Outer Space Explorations, 3. Digital Media, 4. Stock Market & Finance, 5. Camera,
6. Automated Cars, 7. Fitness Apps, 8. Facial Recognition, 9. Voice Auxiliaries, 10. Healthcare

2) Define Artificial Intelligence

Ans. Artificial Intelligence refers to systems or machines that mimic human intelligence to
perform tasks and can improve themselves based on the information they collect.

AI is also the field of study that develops these intelligent machines. This field explores
techniques like machine learning and neural networks to achieve machine intelligence.

3) How does the Anomaly Detector Service evaluate real-time data for anomalies?

Ans. The Anomaly Detector service in Azure (or similar services) uses a two-pronged
approach to evaluate real-time data for anomalies:

1. Statistical Modeling: It creates a statistical model of your time series data. This model
captures the typical patterns and variations within the data.
2. Real-time Anomaly Scoring: As new data points arrive, the service compares them to
the established model. If a data point deviates significantly from the expected range
predicted by the model, it's flagged as a potential anomaly.

4) What is Azure Automated Machine Learning?

Ans. Azure Automated Machine Learning (AutoML) is a cloud-based service offered by

Microsoft that simplifies the process of building machine learning models.

Traditionally, building machine learning models involves a complex workflow with multiple
stages. AutoML automates many of these steps, making it faster and easier to develop effective
models, especially for those who may not have extensive machine learning expertise.

5) What is common flow of activity in Power BI?

Ans. The common flow of activity in Power BI involves following points:

1. Get Data: This is the starting point, where you connect Power BI to various data
sources. It can be Excel spreadsheets, cloud databases, on-premises databases, web
services, or even social media platforms. Power BI offers a wide range of connectors to
facilitate data import.
2. Transform Data: Once imported, the data might need some cleaning and
transformation. Power BI provides a user-friendly interface for data cleaning tasks like
handling missing values, correcting data types, and filtering out irrelevant information.
You can also use Power Query Editor for more advanced data manipulation.
 6) Define Machine Learning.

Ans. Machine learning is a branch of artificial intelligence (AI) and computer science which
focuses on the use of data and algorithms to imitate the way that humans learn, gradually
improving its accuracy.

Machine Learning is the general term for when computers learn from Data.

7) What does Data Granularity mean?

Ans. Data granularity refers to the level of detail or precision within a dataset.
It essentially reflects how "fine-grained" or "coarse-grained" your data is.

8) Define Data. Give an example.

Ans. Data is essentially information that can be collected, organized, and analyzed to reveal
patterns, trends, and relationships. It can be anything from numbers and statistics to text,
images, and even audio or video.

Here's an example to illustrate:

Imagine you run a small bakery. Data you might collect includes:

 Sales figures: The number of croissants, cookies, and other pastries sold each day.
(This is numerical data)
 Customer feedback: Comments from customers about their experience, their favorite
items, or suggestions for improvement. (This is textual data)
 Ingredient inventory: The amount of flour, sugar, butter, and other ingredients you
have in stock. (This can be numerical or categorical data)

By analyzing this data, you can gain valuable insights. For instance, sales figures might reveal
which pastries are most popular, while customer feedback can help you identify areas for
improvement. Inventory data can help you ensure you have enough supplies to meet customer
demand and avoid stockouts.

SECTION - B

Answer any 4 questions. Each carries 5 marks. (4 x 5 = 20)

9) Write a note on the Challenges and Risks of Artificial Intelligence.

Ans: Challenges and Risks of Artificial Intelligence

While Artificial Intelligence (AI) holds immense potential for progress, it also presents
significant challenges and risks that need careful consideration. Here's a closer look at some
key areas of concern:

Bias and Fairness: AI algorithms are trained on data sets created by humans, and these data
sets can reflect societal biases. This can lead to discriminatory outcomes in areas like loan
approvals, hiring practices, or criminal justice systems.

Explainability and Transparency: Many AI systems, especially complex deep learning

models, are opaque in their decision-making processes. This lack of transparency makes it
difficult to understand how an AI arrives at a particular conclusion, hindering accountability
and trust.
Job Displacement: Automation powered by AI is likely to displace some jobs, particularly
those involving repetitive tasks. While new jobs will be created in AI fields, there's a risk of a
widening gap between those with the skills to thrive in the AI economy and those who are left
behind.

Security and Privacy: As AI becomes more sophisticated, it could be used for malicious
purposes like cyberattacks or manipulating social media to influence elections. Additionally,
AI systems that collect and analyze vast amounts of personal data raise privacy concerns.

Existential Risks: Some experts warn about the potential for advanced AI to surpass human
intelligence and potentially become a threat to humanity. While this might seem like science
fiction, it's a consideration for responsible development of AI.

10) How Machine Learning' works? Explain with the help of a 'flow chart'.

Ans. Here's a breakdown of how machine learning works visualized as a flowchart:

1. Data Collection: The process begins by gathering data relevant to the task at hand.
This data can come from various sources like sensors, databases, surveys, or text
documents.
2. Data Preprocessing: The collected data might be messy and require cleaning and
preparation before it can be used for training. This could involve handling missing
values, formatting inconsistencies, and removing irrelevant information.
3. Model Selection: Depending on the task (classification, regression, etc.), an
appropriate machine learning algorithm is chosen. Common algorithms include
decision trees, linear regression, and support vector machines.
4. Model Training: The training set is fed into the chosen algorithm. The algorithm
iteratively learns from the data, adjusting its internal parameters to identify patterns and
relationships that map the input data to the desired output. This stage often involves,
Feature Engineering, which means Creating new features from existing data that
might improve the model's performance.
5. Model Evaluation: The trained model is tested on the unseen data from the testing set.
Metrics like accuracy, precision, and recall are used to evaluate how well the model
performs in predicting the desired outcome for new data.
6. Model Deployment (Optional): The final, well-performing model is deployed for real-
world use. This could involve integrating it into a software application, using it to make
predictions on new data streams, or automating tasks based on its predictions.
 11) Explain the tasks that comes under Computer Vision.

Ans. Computer vision encompasses a wide range of tasks that involve computers deriving
understanding from digital images and videos. Here are some of the common tasks:

1. Image Classification: This is the fundamental task of classifying an image into a predefined
category.

 For example, an image classification model could identify whether an image contains a
cat, a dog, or a car.
 Applications include image organization, content moderation, and autonomous vehicle
object detection.

2. Object Detection: This builds upon image classification by not only identifying objects but
also pinpointing their location within an image using bounding boxes or other localization
techniques.

 Object detection is crucial for tasks like self-driving cars that need to identify and react
to pedestrians, vehicles, and traffic signs.
 It's also used in video surveillance and image analysis for security purposes.

3. Image Segmentation: This task goes beyond object detection by dividing an image into
segments that correspond to different objects or regions of interest.

 It can be further classified into:

o Semantic Segmentation: Assigns each pixel in the image to a specific category
(e.g., sky, road, person).
o Instance Segmentation: Identifies and segments individual instances of objects
even if they belong to the same category (e.g., segmenting two different cats in
an image).
 Applications of segmentation include medical image analysis, self-driving car lane
detection, and robot grasping tasks.

4. Image Restoration:This task deals with improving the quality of images that might be
corrupted by noise, blur, or artifacts.

 It can involve techniques like denoising, deblurring, and inpainting (filling in missing
parts of an image).
 Image restoration is used in photo editing applications, recovering old photographs, and
enhancing medical scans.

5. Image Generation: This advanced task involves creating new images based on existing data
or learned patterns.

 Techniques like Generative Adversarial Networks (GANs) are used to generate

photorealistic images, create artistic variations of existing images, or even complete
partially missing images.
 Applications include creating product mockups, generating variations for data
augmentation in training other computer vision models, and artistic exploration.

6. Video Analysis: Computer vision can also be applied to analyze videos, extracting
information beyond single images.

 This can involve tasks like object tracking (following the movement of objects in a
video), activity recognition (identifying actions like walking, running, or jumping), and
event detection (recognizing specific events happening in a video).
 Video analysis is used in video surveillance, traffic monitoring, and sports analytics.
 12) Explain the NLP services available in Microsoft Azure.

Ans. Microsoft Azure provides a suite of Natural Language Processing (NLP) services through
its Cognitive Services platform. These services enable developers to build applications that can
understand, interpret, and generate human language. Here are the key NLP services available in
Azure:

1. Text Analytics : The Text Analytics API offers various NLP capabilities for analyzing text.

 Sentiment Analysis: Determines the sentiment expressed in a piece of text, categorizing it as

positive, neutral, or negative.
 Key Phrase Extraction: Identifies the most important phrases in a text, helping to summarize
the main points.
 Language Detection: Detects the language of the input text.
 Named Entity Recognition (NER): Identifies and categorizes entities (e.g., people,
organizations, locations) within the text.

2. Translator: The Translator service provides real-time translation for multiple languages.

 Text Translation: Translates text from one language to another.

 Speech Translation: Translates spoken language in real-time.
 Language Detection: Detects the source language automatically before translation.

3. Language Understanding (LUIS): LUIS is a cloud-based service that helps build natural
language understanding into apps, bots, and IoT devices.

 Intent Recognition: Identifies the intent behind a user's input (e.g., booking a flight, checking
the weather).
 Entity Extraction: Extracts relevant information from the text that is necessary for fulfilling the
intent.
 Prebuilt Domains: Provides prebuilt models for common domains such as home automation,
communication, and entertainment.

4. Azure Bot Service: Azure Bot Service, integrated with Bot Framework, allows the
development and deployment of intelligent bots.

 Natural Language Interaction: Supports natural language understanding to enable more

human-like interactions.
 Integration with LUIS: Leverages LUIS for advanced intent and entity recognition.
 Channel Support: Deploy bots across multiple channels, including Microsoft Teams, Facebook
Messenger, and Slack.

5. QnA Maker: QnA Maker is a cloud-based API service that creates a question-and-answer
layer over your data.

 Knowledge Base Creation: Converts unstructured data like FAQs into a knowledge base that
can be queried.
 Customizable Responses: Allows customization of responses to improve user interaction.
 Integration: Easily integrates with bots and applications for enhanced customer service.

6. Speech Service: The Speech Service combines speech-to-text, text-to-speech, and speech
translation capabilities.

 Speech-to-Text: Converts spoken language into text.

 Text-to-Speech: Converts text into natural-sounding speech.
 Speech Translation: Provides real-time translation of spoken language.
7. Form Recognizer: Form Recognizer is an AI-powered document extraction service.

 Form Processing: Extracts key-value pairs, tables, and text from forms and documents.
 Custom Models: Allows training of custom models to extract specific data fields from forms.
 Prebuilt Models: Offers prebuilt models for common document types like receipts and
invoices.

8. Custom Vision : Although primarily an image analysis tool, Custom Vision can be combined
with NLP for more comprehensive applications.

 Image Classification: Tags images with labels.

 Object Detection: Identifies and locates objects within images.
 Integration: Can be integrated with NLP services for multimodal applications, such as
describing images in natural language.

9. Text-to-Speech : Provides advanced text-to-speech capabilities for generating human-like

speech from text.

 Neural Voices: Uses neural networks to produce natural and expressive speech.
 Custom Voice: Allows the creation of a unique voice for branding purposes.

These NLP services provided by Microsoft Azure enable developers to incorporate advanced
language processing capabilities into their applications, enhancing user experience and
automating complex tasks.

13) What is Conversational AL? Explain Chatbot.

Ans. Conversational AI refers to technologies that enable computers to understand, process, and
respond to human language in a natural and meaningful way. It encompasses a range of applications,
including chatbots, virtual assistants, and voice-controlled devices, designed to interact with users
through text or speech.

Key Features of Conversational AI:

 Natural Language Understanding (NLU): This involves the ability of the AI to

interpret the meaning behind user utterances, even if phrased differently or with
grammatical errors.
 Dialogue Management: The AI tracks the conversation flow, remembers previous
interactions, and uses this context to generate relevant responses.
 Speech Recognition: Enables communication through spoken language, allowing users
to interact with the AI through voice commands or dictation.
 Text-to-Speech: Enables to convert written text to speech, making it easier to read any
document, by way of hearing it and resulting in saving of time for humans.

Chatbots

Chatbots are a specific application of conversational AI designed to simulate human

conversation. They can engage in dialogue with users, answer questions, and perform tasks
based on the interactions.

Types of Chatbots: There are two main types of chatbots.

 Rule-based Chatbots: These rely on predefined rules and decision trees to respond to
user queries. They are limited in their flexibility and struggle with complex or
unexpected questions.
  Machine Learning Chatbots: These leverage machine learning models to understand
user intent and generate more nuanced responses based on the context of the
conversation. They can continuously learn and improve with more data and interaction.

Use Cases of Chatbots:

1. Customer Support:
o 24/7 Availability: Provide instant support to customers at any time.
o Common Inquiries: Answer frequently asked questions, guide users through
processes, and handle basic issues.
2. E-commerce:
o Product Recommendations: Suggest products based on user preferences and
browsing history.
o Order Tracking: Provide updates on order status and shipping information.
3. Healthcare:
o Symptom Checking: Assist users in assessing symptoms and suggesting possible
conditions.
o Appointment Scheduling: Help patients book appointments with healthcare
providers.
4. Banking and Finance:
o Account Management: Assist users in managing their accounts, checking balances,
and making transactions.
o Financial Advice: Provide personalized financial advice and investment suggestions.
5. Education:
o Tutoring: Offer help with homework and provide explanations of educational content.
o Administrative Assistance: Answer questions about admissions, courses, and campus
events.

Example of a Chatbot Interaction:

User: "What's the weather like today?"

Chatbot: "The weather today is sunny with a high of 75°F and a low of 60°F."

User: "Can you help me book a flight to New York?"

Chatbot: "Sure, when would you like to travel?"

User: "Tomorrow."

Chatbot: "I found a few options for you. The first flight is at 8:00 AM, and the second one is at
3:00 PM. Which one would you like to book?"

Advantages of Chatbots:

1. Efficiency: Can handle multiple interactions simultaneously, providing quick responses.

2. Cost-Effective: Reduce the need for extensive human customer service teams.
3. Consistency: Provide consistent answers and service quality.
4. Scalability: Easily scale to handle increasing volumes of interactions.

Challenges:

1. Complexity of Queries: Handling complex, multi-turn conversations can be challenging.

2. Accuracy: Ensuring the chatbot understands and responds accurately to user input.
3. User Experience: Designing chatbots that provide a natural and engaging user experience.
 14) Explain the basic building blocks in power BI.

Ans. Power BI utilizes several fundamental building blocks that work together to transform
raw data into clear and insightful visuals for data analysis. Here's a breakdown of these
essential elements:

1. Datasets:

 This is the foundation of your analysis in Power BI. Datasets represent the collection of
data you import from various sources.
 Power BI connects to a wide range of data sources, including Excel spreadsheets, cloud
databases, on-premises databases, web services, and even social media platforms.
 Once imported, the data might need some cleaning and transformation within Power BI
to ensure it's in a suitable format for analysis.

2. Data Modeling:

 This stage involves establishing relationships between your imported datasets. These
relationships define how data tables from different sources connect and allow Power BI
to perform insightful analysis across the combined dataset.
 Imagine you have separate tables for sales data and customer information. Data
modeling would define the relationship between the customer ID in the sales table and
the customer ID in the customer information table, enabling you to analyze sales trends
by customer demographics.

3. Measures:

 Measures are calculations you create to summarize your data and derive meaningful
insights. These can be simple aggregations like sum, average, or minimum/maximum,
or more complex calculations using DAX (Data Analysis Expressions) formulas.
 For instance, a sales measure could calculate the total revenue for a specific product
category.

4. Dimensions:

 Dimensions represent the categories or attributes within your data that you use to filter,
slice, and dice your analysis. They act as the filters and groupings for your
visualizations.
 Examples of dimensions include time periods (year, quarter, month), product
categories, customer segments, or geographic locations.

5. Visualizations:

 This is where Power BI truly shines! Here you create reports using the transformed and
modeled data. Power BI offers a rich library of charts, graphs, and tables to represent
your data insights visually.
 You can drag-and-drop data fields onto the canvas, customize chart types, and format
visuals to effectively communicate your findings. Common visualizations include bar
charts, line charts, pie charts, maps, and scatter plots.

6. Reports & Dashboards:

 Reports serve as the canvas where you arrange your visualizations to tell a data story.
You can combine various charts and tables to present a comprehensive analysis of a
specific topic.
 Dashboards are single-page summaries that consolidate key metrics and visualizations
from multiple reports. They provide a high-level overview of key performance
 indicators (KPIs) and are ideal for monitoring trends or sharing insights with
stakeholders.

7. DAX (Data Analysis Expressions):

 DAX is a formula language within Power BI that allows you to create custom
calculations and extend the capabilities of measures and calculated columns.
 With DAX, you can perform complex calculations, create advanced filtering logic, and
derive deeper insights from your data.

Understanding these building blocks and how they work together is essential for effectively
using Power BI to transform data into actionable insights.