1.

Artificial Intelligence (AI)

Definition

Artificial Intelligence is a branch of computer science that focuses on designing intelligent agents capable of performing
tasks such as learning, reasoning, problem solving, and decision making similar to humans.

Types of AI (Based on Capability)

Narrow AI (Weak AI)

• Designed to perform only one specific task efficiently.

• Works within a limited predefined domain.

• Cannot transfer knowledge to other tasks.

• Most practical AI systems belong to this type.

• Operates using fixed rules or trained models.

Example: • Google Assistant

• Chatbots
• Face recognition systems

General AI (Strong AI)

• Capable of performing any intellectual task like a human.

• Can reason, learn, and adapt to new situations.

• Possesses general problem-solving ability.

• Can apply knowledge across different domains.

• Exists only as a theoretical concept.

Example: Human-like robot (hypothetical).

Super AI

• Surpasses human intelligence in all aspects.

• Capable of self-learning and self-improvement.

• Can make independent decisions better than humans.

• May possess emotions and creativity.

• Considered a future concept in AI research.

Example: Imaginary advanced AI systems, Hypothetical future AI systems
2. Problem and Problem Formulation in AI

Definition of Problem

A problem in AI is defined as a situation where an agent must choose a sequence of actions to move from an initial state
to a desired goal state.

Problem Formulation

• Initial state describes the starting position of the agent.

• Actions define all possible moves available to the agent.

• Transition model explains the result of each action.

• Goal test checks whether the goal state is achieved.

• Path cost measures the cost of reaching the goal.

Example: In the vacuum cleaner problem, the goal is to clean all rooms.

3. Breadth First Search (BFS) and Depth First Search (DFS)

Breadth First Search (BFS)

• BFS explores nodes level by level in the search tree.

• It uses a queue data structure for traversal.

• BFS always finds the shortest path to the goal.

• It is complete and optimal for unweighted graphs.

• Requires large memory for storing nodes.

Example: Level order traversal of a binary tree.

Depth First Search (DFS)

• DFS explores nodes deeply before backtracking.

• It uses stack or recursion for implementation.

• DFS requires less memory than BFS.

• It may not find the shortest path.

• DFS may get stuck in infinite depth.

Example: Preorder traversal of a binary tree.
4. Iterative Deepening Depth First Search (IDDFS)

Definition

IDDFS is a search strategy that repeatedly applies DFS with increasing depth limits until the goal is found.

Key Points

• Combines completeness of BFS and memory efficiency of DFS.

• Searches depth 0 first, then depth 1, then deeper levels.

• Uses limited memory compared to BFS.

• Guarantees optimal solution in unweighted graphs.

• Suitable for large search spaces.

Example: Searching a goal node at depth 3 using depth limits 0,1,2,3.

5. 8-Puzzle Heuristic Search Strategies

Definition

The 8-puzzle is a sliding tile problem used to demonstrate heuristic search techniques in AI.

State Space Representation

• Each arrangement of tiles represents a unique state.

• Operators move the blank tile left, right, up, or down.

• Initial state is the starting arrangement of tiles.

• Goal state is the desired tile arrangement.

• Path cost is the number of moves performed.

Heuristic Search
• Heuristic search uses domain knowledge to guide the search efficiently.

• A heuristic function estimates the cost to reach the goal.

• It reduces the number of explored states.

• Improves search performance compared to uninformed search.

• Commonly used with A* algorithm.

6. Write a Note on Constraint Satisfaction Problem (CSP)

Definition

A Constraint Satisfaction Problem (CSP) is a problem in which a set of variables must be assigned values from their
respective domains while satisfying a set of constraints.

Components of CSP

• Variables represent problem entities to be assigned values.

• Domains define possible values for each variable.

• Constraints restrict the combinations of values.

• Constraints may be unary, binary, or higher order.

• A solution satisfies all constraints simultaneously.

CSP Characteristics

• CSPs are represented using constraint graphs.

• Solutions are found through systematic search.

• Backtracking is commonly used.

• Heuristics improve efficiency.

• CSPs may have one or multiple solutions.

7. Approaches to Knowledge Representation

Definition

Knowledge representation is the method of storing and organizing knowledge so that AI systems can reason and infer
new information.

Approaches

• Logical representation uses formal logical expressions to represent facts and relationships and supports
reasoning through inference.

• Semantic networks represent knowledge in the form of graphs where nodes denote concepts and links denote
relationships.

• Frames organize knowledge into structured units called frames consisting of slots and corresponding values.

• Production rules represent knowledge using condition–action pairs in the IF–THEN format.

• Ontologies define shared and reusable domain concepts along with their relationships in a formal manner.

• Scripts represent knowledge as a sequence of events for stereotyped situations.

• Conceptual dependency represents meaning of sentences using primitive actions and relationships.

Example:- Patient visits doctor → reports fever → diagnosis → flu prescribed

8. Propositional Logic and First-Order Logic

Propositional Logic

• Uses simple statements that are either true or false.

• Does not use variables or quantifiers.

• Uses logical connectives like AND, OR, NOT.

• Easy to implement and understand.

• Limited expressive power.

Example: If it rains, the road is wet.

First-Order Logic

• Uses predicates to represent relations.

• Uses variables and quantifiers.

• Represents objects and their properties.

• More expressive than propositional logic.

• Widely used in knowledge-based systems.

Example: All humans are mortal.

9. Knowledge Representation Using Frames

Definition

Frames are structured data units used to represent stereotyped objects and situations.

Key Points

• Frames consist of slots and slot values.

• Slots store properties of objects.

• Frames support inheritance of attributes.

• Default values can be assigned to slots.

• Similar to object-oriented representation.

Example: Student frame with name, course, CGPA.
10. Semantic Networks

Definition

Semantic networks represent knowledge using a graph structure with nodes and links.

Key Points

• Nodes represent objects or concepts.

• Links represent relationships between concepts.

• Supports inheritance mechanism.

• Easy to visualize and understand knowledge.

• Commonly used in expert systems.

Example: Dog is-an Animal.

11. Ontologies

Definition

Ontology is a formal and explicit specification of concepts, relationships, and rules in a particular domain.

Characteristics

• Represents knowledge formally and clearly.

• Shared and reusable across systems.

• Machine-readable representation.

• Domain-specific knowledge modeling.

• Supports reasoning and inference.

Applications

• Used in semantic web technologies.

• Applied in healthcare and medical systems.

• Used in natural language processing.

• Supports intelligent information retrieval.

• Helps in knowledge management systems.

Example: Medical ontology with disease and treatment.