1. What is reinforcement learning?

Reinforcement learning is a type of machine learning where an agent learns to make

decisions by trial and error in an environment, aiming to maximize cumulative rewards. The agent
interacts with the environment and learns through the concept of feedback, where it receives
rewards or penalties based on its actions. By using a trial-and-error approach, the agent gradually
learns the best actions to take in different situations to maximize the cumulative reward.

2. Explain BFS,

Breadth-first search is a graph traversal algorithm that starts traversing the graph from the
root node and explores all the neighboring nodes. Then, it selects the nearest node and
explores all the unexplored nodes. While using BFS for traversal, any node in the graph can
be considered as the root node.

There are many ways to traverse the graph, but among them, BFS is the most commonly
used approach. It is a recursive algorithm to search all the vertices of a tree or graph data
structure. BFS puts every vertex of the graph into two categories - visited and non-visited. It
selects a single node in a graph and, after that, visits all the nodes adjacent to the selected
node.

Applications of BFS algorithm

The applications of breadth-first-algorithm are given as follows -

○ BFS can be used to find the neighboring locations from a given source location.

○ In a peer-to-peer network, BFS algorithm can be used as a traversal method to find

all the neighboring nodes. Most torrent clients, such as BitTorrent, uTorrent, etc.
employ this process to find "seeds" and "peers" in the network.

○ BFS can be used in web crawlers to create web page indexes. It is one of the main
algorithms that can be used to index web pages. It starts traversing from the source
page and follows the links associated with the page. Here, every web page is
considered as a node in the graph.

○ BFS is used to determine the shortest path and minimum spanning tree.
 ○ BFS is also used in Cheney's technique to duplicate the garbage collection.

○ It can be used in ford-Fulkerson method to compute the maximum flow in a flow

network.

Algorithm

The steps involved in the BFS algorithm to explore a graph are given as follows -

Step 1: SET STATUS = 1 (ready state) for each node in G

Step 2: Enqueue the starting node A and set its STATUS = 2 (waiting state)

Step 3: Repeat Steps 4 and 5 until QUEUE is empty

Step 4: Dequeue a node N. Process it and set its STATUS = 3 (processed state).

Step 5: Enqueue all the neighbours of N that are in the ready state (whose STATUS = 1) and
set

their STATUS = 2

(waiting state)

[END OF LOOP]

Step 6: EXIT

DFS
It is a recursive algorithm to search all the vertices of a tree data structure or a graph. The
depth-first search (DFS) algorithm starts with the initial node of graph G and goes deeper
until we find the goal node or the node with no children.

Because of the recursive nature, stack data structure can be used to implement the DFS
algorithm. The process of implementing the DFS is similar to the BFS algorithm.

1. Step 1: SET STATUS = 1 (ready state) for each node in G

2. Step 2: Push the starting node A on the stack and set its STATUS = 2 (waiting state)

3. Step 3: Repeat Steps 4 and 5 until STACK is empty

4. Step 4: Pop the top node N. Process it and set its STATUS = 3 (processed state)

5. Step 5: Push on the stack all the neighbors of N that are in the ready state (whose
STATUS = 1) and set their STATUS = 2 (waiting state)

6. [END OF LOOP]

7. Step 6: EXIT

Applications of DFS algorithm

○ DFS algorithm can be used to implement the topological sorting.

○ It can be used to find the paths between two vertices.

○ It can also be used to detect cycles in the graph.

○ DFS algorithm is also used for one solution puzzles.

○ DFS is used to determine if a graph is bipartite or not.

3. Explain A* algorithm with example.

A * algorithm is a searching algorithm that searches for the shortest path between the initial and the
final state. It is used in various applications, such as maps.

In maps the A* algorithm is used to calculate the shortest distance between the source (initial state) and
the destination (final state).

Example

A* algorithm is very useful in graph traversals as well. In the following slides, you will see how the
algorithm moves to reach its goal state.

Suppose you have the following graph and you apply A* algorithm on it. The initial node is A and the goal
node is E.

To explain the traversal of the graph above, check out the slides below.

At every step, the f-value is being re-calculated by adding together the g and h values. The minimum
f-value node is selected to reach the goal state. Notice how node B is never visited.

4. Write down various Knowledge Representation issues.

5. Forward vs Backward Reasoning.

6. Explain Inductive and Deductive Learning.

7. Short note on Predicate Logic.
Predicate logic, also known as first-order logic, is a formal system used in mathematics,
philosophy, linguistics, and computer science to represent and reason about propositions
involving quantified variables. It extends propositional logic by introducing quantifiers and
predicates, enabling the representation of more complex statements and the expression of
relationships between objects. Here's a brief overview of predicate logic:

​ Components:
● Predicates: Express properties or relations that can be true or false, such as
"is a mammal" or "is greater than."
● Quantifiers: Include "for all" (∀) and "there exists" (∃), used to express
statements about all or some elements in a domain.
● Variables: Represent elements in the domain of discourse, and they can be
universally or existentially quantified.
​ Syntax:
● Atomic formulas are built from predicates applied to terms or variables.
● Complex formulas are constructed using logical connectives such as
conjunction (∧), disjunction (∨), implication (→), and negation (¬).
​ Expressiveness:
● Predicate logic enables the precise representation of complex statements
involving relationships, properties, and quantified variables.
● It allows for the formalization of mathematical and philosophical arguments
and the expression of intricate logical relations.
​ Applications:
 ● In Mathematics: Predicate logic is used in the formalization of mathematical
theories, proofs, and theorems.
● In Computer Science: It is used in the specification of programming
languages, formal verification, and automated theorem proving.
● In Philosophy: It is employed to represent philosophical arguments and
reason about concepts such as existence, identity, and properties.
​ Limitations:
● Predicate logic does not capture the full complexity of natural language,
leading to challenges in accurately representing some linguistic constructs
and contexts.
● It can become unwieldy and complex when dealing with highly intricate and
nested logical structures.

8. What is semantic network-based representation?

9. Explain NLP.
Natural languages are languages that living creatures use for communication,
A machine is considered to be really intelligent only when it can understand and interpret or speak
the matter of natural language.
The capability to understand, interpret and communicate through natural language is a very
important criteria of intelligent behaviour.

10. Explain ANN architecture in detail.

11. A man is known to speak the truth 4 out of 5 times. He throws a die and reports that the
number obtained is a four. Find the probability that the number obtained is a four using
Bayes Theorem.
12. Explain Alpha-beta and Minimax Algorithm, how they are used to find best path.

​ Minimax Algorithm:
● The Minimax algorithm is a decision-making algorithm used in two-player,
zero-sum games to minimize the maximum possible loss. It involves
recursively evaluating all possible moves from the current state of the game
and choosing the move that leads to the best possible outcome.
 ● The algorithm operates by assuming that the opponent is also playing
optimally and attempting to minimize the player's potential gains.
● It uses a tree-like search, evaluating each possible move and its
consequences, and then backtracking to select the move that leads to the best
possible outcome.
● The goal is to maximize the player's chances of winning while considering the
worst-case scenario.
​ Alpha-Beta Pruning:
● Alpha-beta pruning is an optimization technique used in conjunction with the
Minimax algorithm to reduce the number of nodes evaluated in the game
tree. It helps in cutting off branches of the game tree that are unlikely to lead
to a better solution than the current best.
● By keeping track of the alpha (the best possible move for the maximizing
player) and beta (the best possible move for the minimizing player), the
algorithm can discard branches that will not affect the final decision.
● Alpha-beta pruning significantly reduces the number of nodes that need to be
evaluated, making the search more efficient without affecting the final
outcome.
​ Role in Finding the Best Path:
● In the context of game playing, both the Minimax algorithm and alpha-beta
pruning are used to search the game tree and determine the best move for a
player, considering the opponent's potential moves.
● These algorithms help in finding the optimal path by systematically exploring
all possible moves and choosing the one that leads to the most favorable
outcome, given the assumption of optimal play by the opponent.
● They are instrumental in game-playing AI systems, enabling computers to
make intelligent and strategic decisions in complex game scenarios.

13. What is the role of machine leaning in AI?

Machine Learning in AI:

● Machine learning plays a critical role in AI by enabling systems to learn from

data and make predictions or decisions without explicit programming.
● It involves the development of algorithms and models that allow computers
to improve their performance on a specific task through learning from
experience.
● Machine learning is used in various AI applications, including natural
language processing, computer vision, speech recognition, and
recommendation systems, among others.
 ● It enables AI systems to adapt to new data and make accurate predictions or
decisions, even in complex and uncertain environments, thereby enhancing
their overall capabilities and performance.

14. Consider the following facts and translate these sentences into formulas in predicate
logic. a. There are at least two bears.
b. Everyone loves everyone except himself.
c. Every student except George smiles.
d. Every boy who loves Mary hates every other boy who Mary loves.
15. How Natural Language Processing could be beneficial in Recommendation System,
explain?

Semantic Understanding: NLP helps in understanding the context and semantics of user queries,
reviews, and product descriptions. This allows recommendation systems to grasp the nuances of
user preferences and provide more relevant and personalized recommendations.

Sentiment Analysis: NLP facilitates sentiment analysis of user reviews and feedback. By
understanding the sentiment expressed in user reviews, recommendation systems can identify the
specific aspects of products that users like or dislike, thereby refining the recommendations based
on user sentiment and feedback.

Content Analysis: NLP enables recommendation systems to analyze the content of product
descriptions, specifications, and user-generated content. This analysis helps in identifying key
features and attributes that are relevant to users, which can then be used to make informed
recommendations based on user preferences and requirements.

User Profiling and Personalization: By analyzing user-generated text, comments, and

interactions, NLP helps in creating more accurate user profiles. These profiles can then be used to
customize and personalize recommendations based on individual preferences, behavior, and
interests, leading to a more engaging and tailored user experience.

Contextual Recommendations: NLP helps in understanding the context of user queries and
interactions. This contextual understanding enables recommendation systems to provide more
contextually relevant recommendations, taking into account the user's current needs, preferences,
and circumstances.

Enhanced Search and Discovery: NLP can improve the search and discovery capabilities of
recommendation systems by enabling more accurate and efficient search based on natural language
queries. This allows users to find relevant products or content more easily and quickly, enhancing
the overall user experience.

16. Three urns contain 6 red, 4 black; 4 red, 6 black, and 5 red, 5 black balls respectively. One
of the urns is selected at random and a ball is drawn from it. If the ball drawn is red, find the
probability that it is drawn from the first urn. Using Bayes Theorem.
17. Explain Neural Network in detail.
18. What are some real-life applications of Artificial Intelligence?
Virtual Assistants: AI-powered virtual assistants like Siri, Alexa, and Google Assistant help users
perform tasks such as setting reminders, answering queries, and controlling smart home devices.

Autonomous Vehicles: AI plays a critical role in the development of self-driving cars, enabling
them to perceive their environment, make decisions, and navigate safely without human
intervention.

Healthcare: AI is used in healthcare for tasks such as medical image analysis, diagnosis, drug
discovery, and personalized treatment planning, leading to improved patient care and outcomes.

Finance and Banking: AI is used for fraud detection, risk assessment, algorithmic trading, and
customer service, helping financial institutions streamline operations and enhance security.

E-commerce and Marketing: AI-driven recommendation systems, personalized marketing

campaigns, and chatbots enhance customer experiences, drive sales, and optimize marketing
strategies.

Manufacturing and Robotics: AI is used in manufacturing for process automation, predictive

maintenance, quality control, and collaborative robots (cobots) that work alongside humans to
perform tasks.

Natural Language Processing (NLP): AI-powered NLP applications are used for sentiment
analysis, language translation, chatbots, and voice recognition systems, improving communication
and language understanding.

Cybersecurity: AI is used to detect and respond to cybersecurity threats, identify anomalies, and
protect systems and networks from malicious activities and attacks.

Education: AI is utilized in adaptive learning platforms, intelligent tutoring systems, and

personalized learning experiences, enabling educators to customize content and provide
personalized guidance to students.

Smart Cities: AI is applied in managing traffic flow, optimizing energy usage, monitoring public
safety, and improving urban infrastructure, contributing to the development of efficient and
sustainable smart cities.
19. What is meant by the term artificial intelligence? How it is different from natural
intelligence?

Artificial Intelligence (AI) refers to the simulation of human intelligence processes by machines,
especially computer systems. These processes include learning, reasoning, problem-solving,
perception, and language understanding. AI aims to create systems that can perform tasks that
would typically require human intelligence. AI systems can analyze large amounts of data, recognize
patterns, and make decisions or predictions based on the data analysis.

On the other hand, natural intelligence refers to the cognitive abilities and mental faculties exhibited
by humans and animals. It involves the capacity for learning, understanding complex concepts,
adapting to new situations, and applying knowledge to solve problems. Natural intelligence
encompasses various cognitive abilities such as reasoning, perception, emotional understanding,
and creativity, which are inherent in biological organisms and have evolved over millions of years.

Here are some key differences between artificial intelligence and natural intelligence:

Origin: Artificial intelligence is created by humans through programming and designing algorithms,
whereas natural intelligence is an inherent feature of living organisms and is the product of
biological evolution.

Limitations: Artificial intelligence is limited to the tasks and functions for which it has been
programmed, while natural intelligence exhibits a broad range of cognitive abilities and can adapt to
various complex situations.

Learning Ability: AI systems can learn from data and improve their performance over time, but
they typically require large amounts of data and training to do so. In contrast, natural intelligence
allows organisms to learn from a variety of experiences, make connections, and apply knowledge in
diverse contexts.

Emotional Understanding and Creativity: Natural intelligence involves emotional understanding,

empathy, and creativity, which are aspects that are currently challenging for artificial intelligence
systems to replicate effectively.

Biological vs. Synthetic: Natural intelligence is a product of biological processes and the
functioning of the human brain, whereas artificial intelligence is based on the computational power
of machines and the application of algorithms and programming techniques.
20. Differentiate between local search and global search

Criteria Local Search Global Search

Find a satisfactory solution, usually close to

Find an optimal solution by exploring the
Goal the current state, without considering the
entire search space systematically.
entire search space.

Iterative improvement on the current Systematic exploration of the entire search

Approach
solution. space.

May find a solution that is satisfactory but Aims to find the optimal solution based on
Solution Quality
not necessarily optimal. specific criteria or constraints.

More complex as it involves exploring the

Less complex as it focuses on a limited set of
Complexity entire search space, which can be
neighboring solutions.
computationally intensive.

Focuses on a limited region or neighborhood Explores the entire search space, considering
Search Space
of the current solution. all possible solutions.
 Hill climbing, Simulated Annealing, and Depth-first search, Breadth-first search, and
Examples
Genetic Algorithms. A* algorithm.

21. Write short notes on Bayesian classifier.

Naive Bayes classifiers are a collection of classification algorithms based on
Bayes’ Theorem. It is not a single algorithm but a family of algorithms where
all of them share a common principle, i.e. every pair of features being
classified is independent of each other.
22. What is the future of Artificial Intelligence?
● Increased automation: AI is already being used to automate many tasks in a variety
of industries, from manufacturing to customer service. In the future, we can expect
to see even more automation, as AI becomes more capable and less expensive.
● More powerful AI chips: The development of new AI chips is making it possible to
build more powerful and efficient AI systems. This is leading to advances in areas
such as natural language processing, image recognition, and machine learning.
● More data: AI is trained on data, so the more data we have, the better AI systems
can become. As we continue to generate more data, we can expect to see even more
powerful and intelligent AI systems in the future.
● More widespread adoption: AI is already being used in a variety of industries, but
it is still in its early stages of adoption. In the future, we can expect to see AI become
more widespread, as businesses and organizations recognize the benefits that it can
offer.

Here are some specific examples of how AI is expected to impact the future:

● Healthcare: AI is already being used to develop new drugs and treatments, diagnose
diseases, and provide personalized care to patients. In the future, we can expect to
see AI play an even greater role in healthcare, helping to improve outcomes and
reduce costs.
● Transportation: AI is already being used to develop self-driving cars and trucks. In
the future, we can expect to see AI-powered transportation become more
widespread, making it safer and more efficient to get around.
● Manufacturing: AI is already being used to automate tasks in factories and
warehouses. In the future, we can expect to see AI play an even greater role in
manufacturing, helping to improve efficiency and quality.
 ● Customer service: AI is already being used to chatbots and other customer service
tools. In the future, we can expect to see AI become more sophisticated, able to
provide more personalized and helpful customer service.

23. What are the types of Artificial Intelligence?

24. How many types of agents are there in Artificial Intelligence?

Simple Reflex Agents: These agents make decisions based on the current situation. They don't
consider the consequences of their actions and only respond to the immediate environment.

Model-Based Reflex Agents: These agents have a built-in model of the world that allows them to
anticipate how their actions will affect the environment. They make decisions based on both the
current situation and their internal model.

Goal-Based Agents: These agents have specific goals to achieve. They evaluate the current
situation, consider different actions, and choose the one that helps them get closer to their goals.

Utility-Based Agents: These agents not only have goals but also assign values to the outcomes.
They consider the possible actions and their potential outcomes, and they choose the action that
leads to the most desirable result.

25. What is Overfitting and Underfitting?

Overfitting: Overfitting happens when a machine learning model tries too hard to learn
from the training data, including the noise and random fluctuations. It ends up fitting the
training data very closely but performs poorly on new, unseen data. It's like memorizing
answers for a test without understanding the concepts, so you might do well on the exact
questions you memorized but poorly on related questions you haven't seen before.

Underfitting: Underfitting happens when a machine learning model is too simple to capture
the underlying patterns in the data. It fails to learn from the training data and performs
poorly on both the training data and new, unseen data. It's like trying to fit a straight line to
a complex, curvy shape – it won't match the data well, no matter how you adjust it.
26. What are the techniques used to avoid overfitting?
1. Increase training data.
2. Reduce model complexity.
3. Early stopping during the training phase (have an eye over the loss
over the training period as soon as loss begins to increase stop
training).
4. Ridge Regularization and Lasso Regularization.
5. Use dropout for neural networks to tackle overfitting.

27. What is the difference between Natural language processing and text mining?

Criteria Natural Language Processing (NLP) Text Mining

Focuses on the interaction between Focuses on deriving high-quality

Definition
computers and human languages. information from text.
 Aims to enable computers to understand, Aims to extract valuable patterns and

Objective interpret, and generate human language knowledge from a large amount of text

in a valuable way. data.

Emphasizes on the processing and Focuses on the extraction of useful

Focus
analysis of natural language text. information from text data.

Involves techniques such as tokenization, Involves techniques such as

Techniques parsing, sentiment analysis, and language information retrieval, text

generation. categorization, and entity extraction.

Applied in tasks such as machine Applied in tasks such as document

Applications translation, sentiment analysis, and clustering, information retrieval, and

speech recognition. text summarization.

Focuses on extracting valuable insights

Emphasis on Focuses on understanding the structure
and knowledge from unstructured text
Structure and meaning of human language.
data.
28. What is Fuzzy logic?

Fuzzy logic is a mathematical logic that deals with reasoning based on degrees of
truth rather than the usual true or false (1 or 0) Boolean logic. It allows for the
modeling of approximate reasoning, where an outcome can be partially true,
partially false, or somewhere in between. Fuzzy logic enables the handling of
imprecise data and ambiguity in decision-making processes, making it particularly
useful in systems where imprecision, uncertainty, and incomplete information are
present.

29. What are some differences between classification and regression?

Discrete and categorical values (classes Continuous values (numeric or real

Output
or labels). numbers).

Predicts the class or category of the input Predicts the output value based on input
Objective
data. features.

Predicting whether an email is spam or Predicting the house price based on its
Example
not. features.

Evaluated using metrics like mean squared

Evaluated using metrics like accuracy,
Evaluation error, mean absolute error, and R-squared
precision, recall, and F1 score.
value.
 Used in tasks such as image recognition,
Used in tasks such as sales forecasting, stock
Use Cases sentiment analysis, and disease
market analysis, and demand prediction.
diagnosis.

Algorithms include decision trees, Algorithms include linear regression,

Algorithms logistic regression, and support vector polynomial regression, and support vector

machines. regression.

Data Type Handles categorical and discrete data. Handles continuous and numerical data.

30. What is an Artificial Neural Network? What are some commonly used Artificial Neural
networks?
31. Suppose you know a farmer. He tells you that despite working hard in the fields, his crop
yield is deteriorating. How can AI help him?
Precision Farming: AI-powered sensors and drones can be used to collect data on soil moisture,
nutrient levels, and crop health. This data can help the farmer make informed decisions about
irrigation, fertilization, and pest control, leading to more efficient resource management and
improved crop yield.

Predictive Analytics: AI can analyze historical and real-time data to predict crop diseases, pests,
and weather patterns. By providing early warnings and recommendations, AI can help the farmer
take proactive measures to protect the crops and minimize losses.

Crop Monitoring and Management: AI-based imaging and computer vision technologies can
monitor crop growth and detect anomalies or stress factors early on. This information can enable
the farmer to take timely actions, such as adjusting irrigation schedules, applying targeted
treatments, or optimizing harvesting times.

Recommendation Systems: AI-driven recommendation systems can suggest the most suitable
crop varieties, planting techniques, and crop rotation strategies based on the farmer's specific soil
and environmental conditions. These recommendations can help the farmer maximize yield and
improve the overall health of the soil.
Data-Driven Decision Making: AI can help the farmer analyze complex datasets, historical trends,
and market demands to make informed decisions about crop selection, pricing strategies, and
market timing. This data-driven approach can optimize the farmer's operational efficiency and
increase profitability.

Automation and Robotics: AI-powered robots and automated machinery can assist in tasks such
as seeding, weeding, and harvesting, reducing manual labor requirements and increasing the overall
productivity of the farm.

32. Customers who bought this also bought this”, you might have seen this when shopping on
Amazon. How do you think this works?
33. What is a Chatbot? How can they help to deliver the best customer support to the
customers?
A chatbot is a computer program designed to simulate human conversation, typically
through text or voice interactions. Chatbots use artificial intelligence (AI) and natural
language processing (NLP) to understand and respond to user queries and requests
in a conversational manner. They can be used in various applications, including
customer support, information retrieval, and task automation. Here's how chatbots
can deliver the best customer support to customers:

● Chatbots offers instant resolutions

● Your business can offer 24/7 customer support
● Chatbots continuously learn
● You can tailor answers to different types of customers
● Your team has more context on each customer
● A consistent user experience is created

34. Suppose you have to explain to a beginner how a face detection system works. How would
you do that?
Looking for Patterns: The system first looks for patterns that match the shape and features of a
human face, like the eyes, nose, and mouth. It does this by checking for specific arrangements of
pixels that typically form a face.

Comparing Patterns: Once the system finds these patterns, it compares them with what it knows
about how faces generally look. It checks if the patterns it found match the patterns it has learned
from many other pictures of faces.

Making a Guess: If the patterns it finds look enough like a face, the system will make a guess and
say, "Hey, I think there's a face here!" It might draw a box around the face to show where it is.
Checking Again: Sometimes the system might get it wrong and think something is a face when it's
not. So, it checks again and tries to make sure it's really a face before it says, "Yes, that's definitely a
face!"

35. If you are starting a new business, how will you use AI to promote your business?

Personalized Marketing Campaigns: Implement AI-powered tools for customer segmentation and
personalized marketing. Utilize customer data to create tailored promotional content, offers, and
recommendations that cater to specific customer preferences and behaviors.

Chatbots and Virtual Assistants: Integrate AI-driven chatbots and virtual assistants on the
business website and social media platforms to provide real-time customer support and assistance.
Use these tools to engage with customers, answer queries, and provide personalized
recommendations.

Predictive Analytics for Customer Behavior: Employ AI-based predictive analytics to anticipate
customer behavior and trends. Use these insights to forecast customer needs, identify market
opportunities, and tailor promotional strategies accordingly.

Social Media Analysis and Sentiment Analysis: Utilize AI tools for social media analysis and
sentiment analysis to monitor customer feedback, reviews, and social media interactions. Use this
information to understand customer perceptions, address concerns, and refine marketing strategies
for improved brand reputation.

Content Creation and Optimization: Utilize AI-powered content creation tools to generate
engaging and relevant content for various promotional channels. Use AI for content optimization,
including SEO strategies, to enhance the visibility and reach of the business across different online
platforms.

Recommendation Systems and Cross-Selling: Implement AI-based recommendation systems to

suggest complementary products or services to customers based on their purchase history and
preferences. Utilize cross-selling strategies to encourage customers to explore additional offerings
and increase their overall engagement with the business.

Data-Driven Decision Making: Leverage AI for data-driven decision-making processes, including

market analysis, competitive intelligence, and pricing strategies. Utilize AI to identify emerging
market trends, assess competitor activities, and optimize pricing models for improved business
performance and competitiveness.