Artificial Intelligence

Course Code : BEEE411L

On completion of this course, the students will be able to

1. Evaluate Artificial Intelligence methods and describe their foundations

2. Apply the principles of AI in solutions that require problem solving, inference, perception knowledge
representation and learning
3. Demonstrate the knowledge of reasoning and representation for solving real world problems
4. Analyze and illustrate search and planning algorithms in problem solving
5. Implement the AI models for Engineering applications

19-12-2024 SELECT VIT- Chennai 1

19-12-2024 SELECT VIT- Chennai 2
19-12-2024 SELECT VIT- Chennai 3
19-12-2024 SELECT VIT- Chennai 4
 Module:1 Agents & Environment

Benefits and risks in AI, AI technique; Agents: Structure, behavior, intelligence,

rationality; Environment: Nature of environment, task environment, properties; Types of
agents: Goal based agents, utility-based agents, learning agents

19-12-2024 SELECT VIT- Chennai 5

19-12-2024 SELECT VIT- Chennai 6
 Artificial Intelligence

“Man Made” “Thinking power”

19-12-2024 SELECT VIT- Chennai 7

19-12-2024 SELECT VIT- Chennai 8
19-12-2024 SELECT VIT- Chennai 9
19-12-2024 SELECT VIT- Chennai 10
19-12-2024 SELECT VIT- Chennai 11
19-12-2024 SELECT VIT- Chennai 12
19-12-2024 SELECT VIT- Chennai 13
19-12-2024 SELECT VIT- Chennai 14
19-12-2024 SELECT VIT- Chennai 15
19-12-2024 SELECT VIT- Chennai 16
19-12-2024 SELECT VIT- Chennai 17
19-12-2024 SELECT VIT- Chennai 18
 AI techniques

•Machine Learning (ML) •Artificial Neural Networks (ANN)

•Deep Learning •Convolutional Neural Networks (CNN)
•Natural Language Processing (NLP) •Recurrent Neural Networks (RNN)
•Computer Vision •Transfer Learning
•Expert Systems •Generative Adversarial Networks (GANs)
•Genetic Algorithms •Bayesian Networks
•Reinforcement Learning •Evolutionary Algorithms
•Fuzzy Logic •Swarm Intelligence
•Speech Recognition •Knowledge Representation
•Robotics and Autonomous Systems •Semantic Analysis

19-12-2024 SELECT VIT- Chennai 19

 AI= Critical Thinking, Problem-Solving Skills, and Decision-Making

19-12-2024 SELECT VIT- Chennai 20

 AI= Critical Thinking, Problem-Solving Skills, and Decision-Making

The human brain's neural network is a complex system of interconnected neurons that send electrical and

chemical signals to each other to process information. Artificial neural networks (ANNs) are software programs
that are inspired by the human brain's neural network and are used to solve problems.

19-12-2024 21
 Benefits and Risks in AI

•Benefits:
• Increased efficiency: Automates repetitive tasks and improves productivity.
• Enhanced decision-making: Processes large datasets to offer insights.
• Real-world applications: Healthcare, finance, education, autonomous systems, etc.
• Accessibility: Improves quality of life (e.g., AI-powered assistive tools).
•Risks:
• Ethical concerns: Bias, data privacy issues, lack of transparency.
• Unemployment: Automation could replace human jobs.
• Security threats: Misuse of AI in cyberattacks or weaponization.
• Lack of control: Risks from autonomous systems behaving unpredictably.

19-12-2024 SELECT VIT- Chennai 22

 Benefits and Risks in AI

Use Case: Healthcare Diagnostics

•Benefit:

• AI-powered systems like IBM Watson

analyze patient symptoms and medical
history to suggest potential diagnoses and
treatments, increasing the accuracy and
speed of medical decision-making.

•Risk:

• A biased algorithm could misdiagnose

conditions for certain demographic groups
due to biased training data, leading to
harmful outcomes.

19-12-2024 SELECT VIT- Chennai 23

 Benefits and Risks in AI
Smart Agriculture:
• AI-driven drones monitor crop health and optimize irrigation. Increased crop yields and reduced water usage.

19-12-2024 SELECT VIT- Chennai 24

 Benefits and Risks in AI

Autonomous Vehicles:
1. Self-driving cars like Tesla use AI to navigate, avoid obstacles, and ensure passenger safety.
2. Benefit: Reduced accidents and increased transportation efficiency.

• Front Camera

• Ultrasonic Sensors

• Aerial

• Computer Software
• LIDAR
•
(Light Detection and
Ranging)

19-12-2024 SELECT VIT- Chennai 25

 AI Technique
AI techniques focus on mimicking human cognitive abilities like problem-solving, learning, and decision-making using
algorithms.

Characteristics of AI techniques:

•Handles large and complex data.

•Generalizes solutions to similar problems.

•Efficient in searching and optimizing.

AI= Critical Thinking, Problem-Solving Skills, and Decision-Making

19-12-2024 SELECT VIT- Chennai 26

 AI Technique
Search Algorithms:
1. Use Case: Google Maps uses algorithms like A* to find the shortest path between locations.
2. Benefit: Efficient navigation for millions of users.
Natural Language Processing (NLP):
1. Use Case: Chatbots like ChatGPT help with customer queries in real-time.
2. Benefit: Saves time and improves customer satisfaction.
Machine Learning (ML):
1. Use Case: Netflix recommends shows based on viewing history and patterns.
2. Benefit: Personalized user experiences.
Computer Vision:
1. Use Case: Facial recognition in security systems.
2. Benefit: Enhanced security with automated monitoring.

19-12-2024 SELECT VIT- Chennai 27

 Agents Agent = Architecture + Agent Program

An agent is anything that perceives its environment through sensors and acts upon it using actuators. Examples include robots,
software bots, or even humans.
Structure of Agents:
• Composed of:
• Percept: Input received from the environment.
• Actuators: Mechanisms to perform actions.
• Agent function: Maps percepts to actions.
Behavior:
• Defined by the agent function.
• Determines how the agent responds to various inputs.
Intelligence:
• An agent is intelligent if it can make decisions based on knowledge, reasoning, or learning.
Rationality:
• An agent is rational if it acts to achieve the best outcome (or expected utility) given the current knowledge and constraints.
19-12-2024 SELECT VIT- Chennai 28
 Types of Agents
Agents differ based on their complexity and design:
1.Goal-Based Agents:
1. Make decisions to achieve specific goals.
2. Use a sequence of actions to achieve desired states.
2. Utility-Based Agents:
1. Use a utility function to evaluate the "desirability" of outcomes.
2. Aim to maximize utility for better decision-making.
3. Learning Agents:
1. Adapt and improve performance based on experience.
2. Components:
1. Learning element: Learns from observations.
2. Performance element: Executes actions.
3. Critic: Provides feedback.
4. Problem generator: Suggests exploratory actions.
19-12-2024 SELECT VIT- Chennai 29
 Types of Agents in AI
1. Simple Reflex Agent

2. Model-Based Reflex Agents

3. Goal-Based Agents

4. Utility-Based Agents

5. Learning Agents

6. Rational Agents

7. Reflex Agents with State

8. Learning Agents with a Model

9. Hierarchical Agents

10. Multi-agent systems

19-12-2024 SELECT VIT- Chennai 30

 Agents
Robotic Vacuum Cleaners:
1. Agent Type: Simple Reflex agent.
2. Description: Cleans the floor by
sensing dirt and avoiding obstacles.
3. Key Feature: Immediate responses
without memory.

Reflex Agent:
1. Use Case: Thermostats adjust
temperature based on room conditions.
2. Feature: No memory or long-term
planning.

19-12-2024 SELECT VIT- Chennai 31

19-12-2024 SELECT VIT- Chennai 32
 Agents
Stock Trading Bots:
1. Agent Type: Goal-based agent.
2. Description: Trades stocks to
maximize profit based on market
trends.
3. Key Feature: Achieves profit
goals with minimal risk.

Goal-Based Agent:
1. Use Case: Delivery robots plan
routes to reach destinations.
2. Feature: Focus on achieving a
specific goal.

19-12-2024 SELECT VIT- Chennai 33

Searching and planning
 Agents
Autonomous Drones:
1. Agent Type: Utility-based agent.
2. Description: Drones deliver packages while
avoiding obstacles.
3. Key Feature: Maximize delivery efficiency
and safety.

Utility-Based Agent:
1. Use Case: Ride-hailing apps find routes
optimizing time and cost.
2. Feature: Decision-making based on utility.

19-12-2024 SELECT VIT- Chennai 35

 Agents
Smart Home Assistants (e.g., Alexa):
1. Agent Type: Learning agent.
2. Description: Learns user preferences
over time to automate tasks like turning
off lights.
3. Key Feature: Adapts to the user’s
lifestyle.

Learning Agent:
1. Use Case: AI tutors adapt to student
learning progress.
2. Feature: Improves with feedback.

19-12-2024 SELECT VIT- Chennai 37

 Environment
The environment refers to the external conditions in which the agent operates.
Nature of Environment:
•Fully observable vs. Partially observable: Can the agent perceive all relevant information?
•Deterministic vs. Stochastic: Is the outcome of actions predictable?
•Episodic vs. Sequential: Are actions independent or dependent on history?
•Static vs. Dynamic: Does the environment change during agent deliberation?
•Discrete vs. Continuous: Is the environment composed of distinct states or continuous?
Task Environment:
•Defines the problem the agent is designed to solve.
•Characterized by the PEAS framework: Performance measure, Environment, Actuators, Sensors.

19-12-2024 38
 Fully observable vs. Partially observable

Deterministic vs. Stochastic

19-12-2024 SELECT VIT- Chennai 39

 Episodic vs. Sequential

Static vs. Dynamic

19-12-2024 SELECT VIT- Chennai 40

 Discrete vs. Continuous

19-12-2024 41