UGC NET Computer Science Artificial Intelligence (AI) Syllabus

Artificial Intelligence (AI) syllabus — full chapter and topic list

Expand any chapter to see its topics and sub-topics. This is the whole examinable outline for Artificial Intelligence (AI) in UGC NET Computer Science, not a summary of it.

1. 1

   Approaches to AI

   6 topics

   • Turing Test and Rational Agent Approaches
   • State Space Representation of Problems
   • Heuristic Search Techniques
   • Game Playing
   • Min-Max Search
   • Alpha Beta Cutoff Procedures
2. 2

   Knowledge Representation

   8 topics

   • Logic
   • Semantic Networks
   • Frames
   • Rules
   • Scripts
   • Conceptual Dependency and Ontologies
   • Expert Systems
   • Handling Uncertainty in Knowledge
3. 3

   Planning

   6 topics

   • Components of a Planning System
   • Linear and Non Linear Planning
   • Goal Stack Planning
   • Hierarchical Planning
   • STRIPS
   • Partial Order Planning
4. 4

   Natural Language Processing

   3 topics

   • Grammar and Language
   • Parsing Techniques
   • Semantic Analysis and Pragmatics
5. 5

   Multi Agent Systems

   7 topics

   • Agents and Objects
   • Agents and Expert Systems
   • Generic Structure of Multiagent System
   • Semantic Web
   • Agent Communication
   • Knowledge Sharing using Ontologies
   • Agent Development Tools
6. 6

   Fuzzy Sets

   8 topics

   • Notion of Fuzziness
   • Membership Functions
   • Fuzzification and Defuzzification
   • Operations on Fuzzy Sets
   • Fuzzy Functions and Linguistic Variables
   • Fuzzy Relations
   • Fuzzy Rules and Fuzzy Inference
   • Fuzzy Control System and Fuzzy Rule Based Systems
7. 7

   Genetic Algorithms (GA)

   4 topics

   • Encoding Strategies
   • Genetic Operators
   • Fitness Functions and GA Cycle
   • Problem Solving using GA
8. 8

   Artificial Neural Networks (ANN)

   7 topics

   • Supervised Learning
   • Unsupervised Learning
   • Reinforcement Learning
   • Single Perceptron
   • Multi Layer Perceptron
   • Self Organizing Maps
   • Hopfield Network

Artificial Intelligence (AI) flashcards for UGC NET Computer Science

18 of 60 cards from the Artificial Intelligence (AI) deck — real questions with worked answers.

1. What is the Turing Test, and what does it aim to determine?

   A test proposed by Alan Turing (1950) in which a human interrogator converses via text with a machine and a human; if the interrogator cannot reliably tell which is the machine, the machine is said to exhibit intelligent (human-like) behaviour. It tests behavioural indistinguishability, not internal reasoning.

2. Contrast the four approaches to AI: acting humanly, thinking humanly, thinking rationally, and acting rationally.

   Acting humanly = Turing Test approach; Thinking humanly = cognitive modelling; Thinking rationally = laws of thought (logic); Acting rationally = the rational agent approach, which acts to achieve the best expected outcome. Modern AI favours the rational agent approach.

3. What is a rational agent?

   An agent that, for each possible percept sequence, selects an action expected to maximize its performance measure, given the evidence from percepts and its built-in knowledge. Rationality depends on the performance measure, prior knowledge, available actions, and percept sequence.

4. What is state space representation of a problem, and what are its components?

   A formulation of a problem as a graph of states. Components: initial state, set of actions/operators, transition (successor) function, goal test, and path cost. A solution is a sequence of actions (path) from the initial state to a goal state.

5. In A* search, what is the evaluation function and the condition for optimality?

   f(n) = g(n) + h(n), where g(n) is the cost from start to n and h(n) is the estimated cost from n to the goal. A* is optimal (with tree search) if h is admissible (never overestimates), and complete/optimal for graph search if h is consistent (monotonic).

6. What distinguishes a heuristic search from an uninformed (blind) search?

   A heuristic (informed) search uses problem-specific knowledge via a heuristic function h(n) to estimate cost to the goal, guiding the search efficiently (e.g., Best-First, A*, Hill Climbing). Uninformed search (e.g., BFS, DFS) uses only the problem definition with no domain guidance.

7. What is an admissible heuristic, and why does it matter?

   A heuristic h(n) is admissible if it never overestimates the true cost to reach the goal, i.e., 0 ≤ h(n) ≤ h*(n). Admissibility guarantees that A* tree search returns an optimal solution.

8. What characterizes the game-playing problems typically studied in AI?

   They are typically two-player, zero-sum, deterministic games of perfect information (e.g., chess, tic-tac-toe), modelled as a game tree where players (MAX and MIN) alternate moves, and one player's gain is the other's loss.

9. How does the Min-Max algorithm choose a move?

   It recursively explores the game tree assuming both players play optimally: MAX nodes take the maximum of children's values, MIN nodes take the minimum, propagating terminal utility values up to the root. MAX then picks the move leading to the highest backed-up value.

10. What is the time complexity of Min-Max, and why is pruning needed?

    Min-Max explores the full game tree with time complexity O(b^d) (b = branching factor, d = depth), which is exponential. Alpha-beta pruning is needed to cut off branches that cannot affect the final decision, reducing effective cost.

11. Explain the alpha and beta values in alpha-beta pruning.

    Alpha = the best (highest) value found so far for MAX along the path; Beta = the best (lowest) value found so far for MIN. Pruning occurs when alpha ≥ beta, since the current branch cannot influence the outcome.

12. What is the best-case complexity of alpha-beta pruning with perfect move ordering?

    With optimal move ordering it examines O(b^(d/2)) nodes instead of O(b^d), effectively halving the depth exponent and allowing search roughly twice as deep in the same time.

13. Distinguish propositional logic from first-order (predicate) logic.

    Propositional logic deals with whole declarative statements (true/false propositions) combined by connectives. First-order/predicate logic adds objects, predicates (relations), functions, and quantifiers (∀, ∃), giving greater expressive power to represent relationships among objects.

14. What is a semantic network in knowledge representation?

    A graphical knowledge representation where nodes represent concepts/objects and labelled directed arcs represent relations between them (e.g., IS-A, HAS-PART). It supports inheritance of properties along IS-A links.

15. What is a frame in knowledge representation?

    A structured representation of a stereotyped object or situation, consisting of slots (attributes) and fillers (values). Slots can hold default values, constraints, or procedures (demons), and frames support inheritance via class hierarchies.

16. What are production rules (Rules) in a knowledge-based system?

    Condition-action pairs of the form IF <condition> THEN <action/conclusion>. A rule-based system uses an inference engine that matches rules against working memory and fires them (forward or backward chaining) to derive conclusions.

17. What is a 'script' in AI knowledge representation (Schank & Abelson)?

    A structure describing a stereotyped sequence of events in a particular context (e.g., the restaurant script). Components include entry conditions, roles, props, scenes, and results, enabling inference of implicit, unstated events.

18. What is Conceptual Dependency (CD) theory?

    A theory (Schank) representing the meaning of sentences in a language-independent way using a small set of primitive acts (e.g., ATRANS, PTRANS, MTRANS, INGEST) so that sentences with the same meaning have the same CD representation.

See more Artificial Intelligence (AI) flashcards →

Planning Artificial Intelligence (AI) for UGC NET Computer Science

Artificial Intelligence (AI) (UGC NET Computer Science) FAQ