AI

2M
1. Explain heuristic function with eg.
Ans:-

Heuristic Function in AI
A heuristic function is an evaluation function that helps guide search algorithms toward
the goal more efficiently by estimating the cost of reaching the goal from a given state. It
takes a problem state as input and returns a numerical value representing its desirability or
estimated cost.
Characteristics of a Heuristic Function:
1. Estimates Path Cost: It provides an estimate of the cost required to reach the goal from
a given node.
2. Guides Search Algorithms: It helps algorithms focus on promising paths, improving
efficiency.
3. Can be Maximization or Minimization: Depending on the problem, a heuristic function
can either maximize (higher cost is better) or minimize (lower cost is better).
4. May Not Always Be Optimal: While heuristics improve efficiency, they don’t guarantee
the best solution in all cases.

Example: 8-Puzzle Problem

The 8-puzzle problem involves arranging numbered tiles in a 3×3 grid to match a goal state,
with one empty tile to move other tiles.
Two Common Heuristic Functions:
1. Misplaced Tiles Heuristic (Hamming Distance, h₁(n))
o Counts the number of misplaced tiles (excluding the blank tile).
o Example: If 8 tiles are in the wrong position, h₁ = 8.
o Admissible Heuristic: Any misplaced tile must move at least once.
2. Manhattan Distance Heuristic (h₂(n))
o Calculates the sum of the horizontal and vertical distances of tiles from their
correct positions.
o Example: If a tile needs to move 2 steps right and 3 steps down, its distance is 2 +
3 = 5.
o h₂ always dominates h₁ (i.e., it provides a more informed estimate).
Since h₂ always gives a better estimate than h₁, it leads to a more efficient search.
2. Explain steps in Problem Formulation with eg.
Ans:-
Steps in Problem Formulation in AI
Problem formulation in AI is the process of defining a problem in a structured way so that an
AI agent can find a solution effectively. It involves defining the state space, actions, goal
state, and path cost.
Key Components of Problem Formulation
A problem can be formally defined using the following five components:
1. Initial State: The starting state of the problem.
2. Actions: The set of all possible actions that can be taken from a given state.
3. Successor Function: Defines the result of applying an action to a state, leading to a new
state.
4. Goal Test: A check to determine if the current state matches the goal state.
5. Path Cost: A function that assigns a numerical cost to a path from the initial state to the
goal state.

Example: 8-Puzzle Problem

The 8-puzzle is a common AI problem where a 3×3 board consists of eight numbered tiles
and one blank tile. The goal is to rearrange the tiles to match the goal state by sliding the
blank tile.
Problem Formulation:
1. Initial State: The board starts with a specific arrangement of tiles, e.g.,

Here, 0 represents the blank space.

2. Actions: The blank tile can move Left, Right, Up, or Down, changing its position
accordingly.
3. Successor Function: If the blank moves down, the new state will be:

4. Goal Test: The state is compared to the desired goal state:

If the current state matches the goal state, the problem is solved.
5. Path Cost: The cost is defined as the number of moves taken to reach the goal state.
3. What are rational agents? Explain with eg.
Ans:-

Rational Agent in AI
A rational agent in Artificial Intelligence (AI) is an entity that makes decisions based on
logical reasoning, knowledge, and perceptual inputs to achieve an optimal outcome.
Rationality in AI means selecting the best action from available options to maximize
performance, even in uncertain environments.

Definition of Rationality and Rational Agent

• Rationality: It refers to the ability of an agent to take actions that maximize its
performance based on knowledge, past experiences, and percept sequences.
• Rational Agent: An agent that perceives the environment, evaluates possible actions, and
chooses the best possible action to achieve its goal efficiently.

Key Factors Determining Rationality

1. Performance Measure – Defines success criteria for the agent.
2. Percept Sequence – The history of perceptions received by the agent.
3. Agent’s Knowledge – Prior knowledge about the environment.
4. Feasible Actions – The actions available for the agent to choose from.
Example of Rational Action

Vacuum Cleaner Agent:

A vacuum cleaner robot in AI is a rational agent. It uses sensors to detect dirt levels, battery
status, and obstacles and then decides its movement efficiently.
• If a room is dirty, the agent chooses to clean it.
• If the battery is low, it navigates to the charging dock before continuing cleaning.
• If it detects an obstacle, it reroutes itself to avoid collisions.
This rational decision-making ensures optimal performance and efficiency in cleaning.
4. Enlist & explain concepts of AI.
Ans:-

1. Machine Learning (ML)

• ML is a subset of AI that enables computers to learn from data and improve performance
without explicit programming.
• Types of ML:
o Supervised Learning: The model is trained on labeled data (e.g., spam detection
in emails).
o Unsupervised Learning: The model identifies patterns in unlabeled data (e.g.,
customer segmentation).
o Reinforcement Learning: The model learns by interacting with the environment
and receiving rewards or penalties (e.g., robotics, game playing).

2. Deep Learning (DL)

• A specialized form of ML that uses artificial neural networks (ANNs) to process large
amounts of data.
• Examples: Image recognition, speech recognition (e.g., Siri, Google Assistant), and self-
driving cars.

3. Natural Language Processing (NLP)

• AI’s ability to understand, interpret, and generate human language.
• Applications: Chatbots, voice assistants, sentiment analysis, language translation.

4. Computer Vision
• AI’s ability to process and analyze visual data (images/videos).
• Used in facial recognition, autonomous vehicles, medical imaging, and object detection.

5. Expert Systems
• AI systems that mimic human decision-making by using a knowledge base and set of
rules.
• Used in medical diagnosis, fraud detection, and business decision-making.

6. Robotics
• AI-driven machines that can perform tasks autonomously or semi-autonomously.
• Examples: Industrial robots, humanoid robots, and robotic process automation (RPA).

7. Fuzzy Logic
• AI technique that deals with reasoning under uncertainty, similar to human decision-
making.
• Used in control systems (e.g., air conditioners, washing machines).
8. Neural Networks
• Computational models inspired by the human brain that process information in layers of
interconnected nodes.
• Forms the foundation of deep learning.

9. Knowledge Representation and Reasoning

• AI’s ability to store knowledge and use it to reason, infer, and make decisions.
• Used in expert systems and semantic web technologies.

10. Genetic Algorithms

• Optimization techniques inspired by natural selection and evolution.
• Used in AI for problem-solving, scheduling, and machine learning model tuning.
5. Explain different definitions of AI.
Ans:-
Artificial Intelligence (AI) has been defined in multiple ways by different researchers,
organizations, and scholars. Here are some key definitions from various perspectives:
1. John McCarthy (1956) – Father of AI
"AI is the science and engineering of making intelligent machines, especially intelligent
computer programs."
• This definition emphasizes AI as a field of study focused on creating machines that can
simulate human intelligence.

2. Alan Turing (1950) – Turing Test Approach

"A machine is considered intelligent if it can mimic human responses so well that a
human evaluator cannot distinguish whether the responses come from a human or a
machine."
• Turing’s approach focuses on AI’s ability to exhibit human-like behavior in
conversations.

3. Russell & Norvig (1995) – AI as an Agent

"AI is the study of agents that perceive the environment and take actions that maximize
their chances of success."
• This definition views AI as an intelligent agent that interacts with its environment to
achieve a goal.

4. The Dartmouth Conference Definition (1956)

"Every aspect of learning or any other feature of intelligence can in principle be so
precisely described that a machine can be made to simulate it."
• This early definition highlights the goal of AI to replicate human cognitive abilities.

5. IBM Definition of AI
"Artificial Intelligence refers to any system that mimics human intelligence to perform
tasks and iteratively improves itself based on collected data."
• Focuses on AI’s self-learning capability using data.

6. Nils Nilsson (1998) – Cognitive Science Approach

"AI is the activity of designing computer systems that perform tasks requiring human
intelligence."
• This emphasizes AI’s role in solving problems that humans typically handle.

7. European Commission (2021)

"Artificial intelligence (AI) refers to software that is developed with techniques such as
machine learning, logic-based approaches, and statistical methods to perform tasks that
normally require human intelligence."
• Highlights AI's reliance on different computational methods to perform intelligent tasks.
6. Differentiate between uninformed search & informed search
techniques.
Ans:-
5M
1. What is PEAS descriptor? Give PEAS descriptor for Part-Picking Robot.
Ans:-

PEAS Descriptor in AI
PEAS stands for Performance measure, Environment, Actuators, and Sensors. It is a
framework used to define the components of an intelligent agent in artificial intelligence
(AI). The PEAS descriptor helps in designing and analyzing agents by specifying their key
characteristics.

PEAS Descriptor for a Part-Picking Robot

A Part-Picking Robot is an AI-based robotic system used in manufacturing and assembly
lines to pick and place parts efficiently. Its PEAS components are:
1. Performance Measure
o Accuracy in picking the correct part
o Speed of picking and placing
o Minimizing errors or misplacements
o Energy efficiency
2. Environment
o Factory or warehouse setting
o Conveyor belts and storage bins
o Varied lighting and temperature conditions
o Presence of humans and other robots
3. Actuators
o Robotic arm for picking and placing
o Grippers or suction mechanisms
o Motors for movement
o Conveyor system for transporting parts
4. Sensors
o Camera or vision system for object detection
o Proximity sensors for obstacle avoidance
o Force sensors for grip control
o RFID or barcode scanners for part identification
This PEAS framework helps in designing and evaluating the effectiveness of a Part-Picking
Robot in an AI-based industrial system.
2. Write short note on Applications of AI.
Ans:-

Applications of AI
Artificial Intelligence (AI) is revolutionizing various industries by enabling machines to
mimic human intelligence. AI applications enhance automation, decision-making, and
efficiency across multiple domains. Below are some key applications:
1. Healthcare
• Medical Diagnosis – AI helps in detecting diseases like cancer and neurological
disorders through medical imaging (e.g., MRI, CT scans).
• Drug Discovery – AI accelerates drug development by predicting molecular interactions.
• Robotic Surgery – AI-powered robots assist in minimally invasive surgeries, improving
precision.
• Virtual Health Assistants – Chatbots and AI assistants provide medical advice and
schedule appointments.

2. Finance & Banking

• Fraud Detection – AI analyzes transaction patterns to identify fraudulent activities.
• Algorithmic Trading – AI executes high-speed financial transactions based on market
trends.
• Credit Scoring & Risk Assessment – AI evaluates loan applicants' financial history to
assess creditworthiness.
• Chatbots in Banking – AI-powered bots assist customers in account management and
queries.

3. Autonomous Systems
• Self-Driving Cars – AI enables vehicles to navigate, recognize objects, and make driving
decisions.
• Drones – AI-powered drones are used in delivery services, military surveillance, and
disaster management.
• Industrial Automation – AI-driven robots handle repetitive tasks in manufacturing,
logistics, and assembly lines.

4. Natural Language Processing (NLP)

• Virtual Assistants – AI powers voice assistants like Siri, Alexa, and Google Assistant.
• Language Translation – AI-driven tools like Google Translate help in multilingual
communication.
• Chatbots & Customer Support – AI-based chatbots handle queries and automate
customer service.
5. Computer Vision
• Facial Recognition – AI is used in biometric security and surveillance.
• Medical Imaging – AI assists in detecting diseases from X-rays, CT scans, and MRIs.
• Object Detection – AI helps in security, agriculture (crop monitoring), and retail (smart
checkouts).

6. E-Commerce & Retail

• Personalized Recommendations – AI suggests products based on user behavior (e.g.,
Amazon, Netflix).
• Inventory Management – AI predicts demand and optimizes stock levels.
• Chatbots & Virtual Shopping Assistants – AI enhances customer experience by
assisting in purchases.

7. Gaming & Entertainment

• AI in Video Games – AI creates realistic NPC behavior and adaptive difficulty levels.
• Content Creation – AI generates music, artworks, and even scripts for movies.
• Deepfake Technology – AI is used in creating hyper-realistic digital content.

8. Cybersecurity
• Threat Detection – AI monitors and detects cyber threats in real time.
• Phishing Prevention – AI identifies fraudulent emails and malicious links.
• Network Security – AI enhances firewalls and intrusion detection systems.

9. Education
• AI Tutors – AI-powered platforms provide personalized learning experiences.
• Automated Grading – AI assists teachers by evaluating assignments and exams.
• Smart Classrooms – AI enables interactive learning using augmented reality (AR) and
virtual reality (VR).

10. Smart Cities & IoT

• Traffic Management – AI optimizes traffic signals and reduces congestion.
• Smart Energy Management – AI predicts energy consumption and improves efficiency.
• Surveillance & Safety – AI-powered cameras detect suspicious activities for enhanced
security
3. What are Intelligent agents? Explain human & robotic agent structure.
Ans:-

Intelligent Agents in AI
An Intelligent Agent (IA) is an entity in Artificial Intelligence (AI) that
perceives its environment through sensors and acts upon it using actuators to
achieve a specific goal. It continuously learns from its environment and makes
decisions to maximize performance.
Types of Intelligent Agents:
1. Simple Reflex Agents – Act only based on the current percept, ignoring
history.
2. Model-Based Reflex Agents – Maintain an internal model of the world to
handle partially observable environments.
3. Goal-Based Agents – Make decisions based on achieving specific goals.
4. Utility-Based Agents – Optimize decisions based on a utility function to
maximize performance.
5. Learning Agents – Continuously learn from experiences and improve
performance over time.

Human Agent Structure:

A human is a natural intelligent agent with the following components:
• Sensors: Eyes, ears, skin, nose, tongue (to perceive the environment).
• Brain (Processing Unit): Analyzes sensory data, makes decisions, and
plans actions.
• Actuators: Hands, legs, vocal cords (used for interacting with the
environment).

Robotic Agent Structure:

A robotic agent is an artificial intelligent agent with:
• Sensors: Cameras, microphones, infrared, ultrasonic, or LIDAR (for
perceiving the surroundings).
• Processor/Controller: AI model, microprocessor, or CPU (for processing
data and making decisions).
• Actuators: Motors, robotic arms, wheels, or speakers (for performing
actions based on decisions).
Both human and robotic agents function similarly but differ in intelligence,
adaptability, and learning capabilities. AI-powered robotic agents can be used
in automation, healthcare, manufacturing, and autonomous systems.
4o
 4. Explain:
o Simple Reflex Agent
o Model-Based Agent
o Goal-Based Agent
o Utility-Based Agent
Ans:-

Types of Intelligent Agents in AI

In Artificial Intelligence (AI), agents perceive their environment through sensors and act upon it
using actuators. Depending on how they make decisions, intelligent agents are categorized into
the following types:

1. Simple Reflex Agent

A Simple Reflex Agent makes decisions based only on the current percept (input from the
environment) and does not consider past experiences or future consequences.
Characteristics:
• Follows a set of predefined condition-action rules (also known as rule-based behavior).
• Does not store past states or learn from experiences.
• Works well in fully observable environments.
Example:
• A thermostat that turns on the heater if the temperature falls below a threshold.
• A traffic light system that changes signals based on real-time sensor input.
Limitations:
• Cannot handle partial observability (i.e., when all data is not available at a given
moment).
• Struggles with complex environments requiring memory and planning.

2. Model-Based Agent
A Model-Based Agent improves upon simple reflex agents by maintaining an internal model of
the environment. This model helps the agent track changes over time and make better
decisions.
Characteristics:
• Stores past perceptions to understand how the environment evolves.
• Uses a model to predict future states and determine actions accordingly.
• Works better in partially observable environments compared to simple reflex agents.
Example:
• A self-driving car that maintains a map of its surroundings and predicts the movement
of other vehicles.
• Robot vacuum cleaners like Roomba, which remember room layouts and obstacles.
Limitations:
• Requires more computational power than simple reflex agents.
 • Needs an accurate model of the environment for optimal performance.

3. Goal-Based Agent
A Goal-Based Agent makes decisions by considering future outcomes and selecting actions that
help achieve a specific goal. Unlike reflex agents, goal-based agents can evaluate different
possibilities before acting.
Characteristics:
• Uses search and planning algorithms to determine the best sequence of actions.
• Can handle more complex tasks requiring long-term decision-making.
• May require significant computation time depending on the problem.
Example:
• A chess-playing AI like AlphaZero, which evaluates different moves to achieve
checkmate.
• Path-planning algorithms in GPS navigation systems that find the best route to a
destination.
Limitations:
• May struggle in dynamic environments where conditions change rapidly.
• Requires a well-defined goal to function effectively.

4. Utility-Based Agent
A Utility-Based Agent extends goal-based agents by not just achieving goals, but also
optimizing performance based on a utility function. This function assigns a numerical value to
different outcomes, allowing the agent to choose the most beneficial action.
Characteristics:
• Uses a utility function to compare different actions.
• Selects the action that maximizes expected benefits.
• Can handle uncertainty and probabilistic environments.
Example:
• Recommendation systems (e.g., Netflix, YouTube) suggest content based on user
preferences to maximize engagement.
• Autonomous trading bots in stock markets select the best investment strategy to
maximize returns.
Limitations:
• Requires accurate utility function design for effective decision-making.
• Can be computationally expensive in complex environments.