2024 question paper

Artificial intelligence

26 questions

  1. Q1a. Which of the following best defines the Turing Test? (i) A method to calculate machine efficiency (ii) A test for machine learning algorithms (iii) A test to determine if a machine can exhibit human-like intelligence (iv) A benchmark for robotic speed20242m

    Module 1: Introduction and Search Techniques

    Which of the following best defines the Turing Test?

    (i) A method to calculate machine efficiency
    (ii) A test for machine learning algorithms
    (iii) A test to determine if a machine can exhibit human-like intelligence
    (iv) A benchmark for robotic speed

    View this question on its own page →
  2. Q1b. The A^* algorithm guarantees optimality if the heuristic used is: (i) Arbitrary (ii) Inconsistent (iii) Admissible and consistent (iv) Random20242m

    Module 1: Introduction and Search Techniques

    The AA^* algorithm guarantees optimality if the heuristic used is:

    (i) Arbitrary
    (ii) Inconsistent
    (iii) Admissible and consistent
    (iv) Random

    View this question on its own page →
  3. Q1c. Which of these search strategies is *not* complete in infinite-depth spaces? (i) Breadth-First Search (ii) Depth-First Search (iii) DFID (iv) A^* Search20242m

    Module 1: Introduction and Search Techniques

    Which of these search strategies is not complete in infinite-depth spaces?

    (i) Breadth-First Search
    (ii) Depth-First Search
    (iii) DFID
    (iv) AA^* Search

    View this question on its own page →
  4. Q1d. In an AND/OR graph, solving an AND node means: (i) Solving one child node is enough (ii) All child nodes must be solved (iii) The node is ignored (iv) Any node can be skipped20242m

    Module 1: Introduction and Search Techniques

    In an AND/OR graph, solving an AND node means:

    (i) Solving one child node is enough
    (ii) All child nodes must be solved
    (iii) The node is ignored
    (iv) Any node can be skipped

    View this question on its own page →
  5. Q1e. Minimax is used in game playing to: (i) Maximize randomness (ii) Minimize evaluation time (iii) Make optimal moves assuming the opponent plays optimally (iv) Store all possible states20242m

    Module 1: Introduction and Search Techniques

    Minimax is used in game playing to:

    (i) Maximize randomness
    (ii) Minimize evaluation time
    (iii) Make optimal moves assuming the opponent plays optimally
    (iv) Store all possible states

    View this question on its own page →
  6. Q1f. Alpha-beta pruning improves Minimax by: (i) Increasing the depth (ii) Ignoring non-optimal branches (iii) Doubling the search time (iv) Reducing the score of nodes20242m

    Module 1: Introduction and Search Techniques

    Alpha-beta pruning improves Minimax by:

    (i) Increasing the depth
    (ii) Ignoring non-optimal branches
    (iii) Doubling the search time
    (iv) Reducing the score of nodes

    View this question on its own page →
  7. Q1g. Propositional logic cannot express which of the following? (i) True/false statements (ii) Compound statements (iii) Logical connectives (iv) Variables and functions20242m

    Module 2: Knowledge Representation and Reasoning

    Propositional logic cannot express which of the following?

    (i) True/false statements
    (ii) Compound statements
    (iii) Logical connectives
    (iv) Variables and functions

    View this question on its own page →
  8. Q1h. Bayesian networks are especially useful in situations with: (i) Complete knowledge and logic (ii) No uncertainty (iii) Deterministic processes (iv) Probabilistic and uncertain information20242m

    Module 2: Knowledge Representation and Reasoning

    Bayesian networks are especially useful in situations with:

    (i) Complete knowledge and logic
    (ii) No uncertainty
    (iii) Deterministic processes
    (iv) Probabilistic and uncertain information

    View this question on its own page →
  9. Q1i. Which of the following is most likely used in deep learning? (i) Decision Trees (ii) Shallow neural networks (iii) Multi-layered neural networks (iv) K-Means algorithm20242m

    Module 3: Machine Learning Fundamentals

    Which of the following is most likely used in deep learning?

    (i) Decision Trees
    (ii) Shallow neural networks
    (iii) Multi-layered neural networks
    (iv) K-Means algorithm

    View this question on its own page →
  10. Q1j. Which of the following is a key step in a genetic algorithm? (i) Regression (ii) Cross-validation (iii) Crossover (iv) Tokenization20242m

    Module 4: Advanced AI Applications

    Which of the following is a key step in a genetic algorithm?

    (i) Regression
    (ii) Cross-validation
    (iii) Crossover
    (iv) Tokenization

    View this question on its own page →
  11. Q2a. A chatbot consistently passes the Turing Test during short conversations but fails at longer ones. What does this imply about the limitations of the test? Suggest an enhancement to the test.20247m

    Module 1: Introduction and Search Techniques

    A chatbot consistently passes the Turing Test during short conversations but fails at longer ones. What does this imply about the limitations of the test? Suggest an enhancement to the test.

    View this question on its own page →
  12. Q2b. Explain how an agent's architecture affects its ability to operate in a partially observable, stochastic environment. Provide a practical example.20247m

    Module 1: Introduction and Search Techniques

    Explain how an agent's architecture affects its ability to operate in a partially observable, stochastic environment. Provide a practical example.

    View this question on its own page →
  13. Q3a. Show that Hill Climbing can get stuck in local maxima, plateaus, and ridges. Suggest at least two variations or techniques to overcome these issues and explain how they help.20247m

    Module 1: Introduction and Search Techniques

    Show that Hill Climbing can get stuck in local maxima, plateaus, and ridges. Suggest at least two variations or techniques to overcome these issues and explain how they help.

    View this question on its own page →
  14. Q3b. Heuristics are designed to guide search algorithms efficiently, but they can sometimes mislead the search. Describe a situation in which A* search performs worse than an uninformed search due to a poorly chosen heuristic.20247m

    Module 1: Introduction and Search Techniques

    Heuristics are designed to guide search algorithms efficiently, but they can sometimes mislead the search. Describe a situation in which A search performs worse than an uninformed search due to a poorly chosen heuristic.*

    View this question on its own page →
  15. Q4a. Define a basic Constraint Satisfaction Problem (CSP) involving three variables and solve it using backtracking. Also explain the role of domain and constraints.20247m

    Module 1: Introduction and Search Techniques

    Define a basic Constraint Satisfaction Problem (CSP) involving three variables and solve it using backtracking. Also explain the role of domain and constraints.

    View this question on its own page →
  16. Q4b. Describe how Particle Swarm Optimization (PSO) works with a simple analogy (e.g., birds flocking). Show how particles update their positions and velocities.20247m

    Module 1: Introduction and Search Techniques

    Describe how Particle Swarm Optimization (PSO) works with a simple analogy (e.g., birds flocking). Show how particles update their positions and velocities.

    View this question on its own page →
  17. Q5a. Create a simple two-player game tree (3 levels) and show how the Minimax algorithm selects the best move for the maximizing player.20247m

    Module 1: Introduction and Search Techniques

    Create a simple two-player game tree (3 levels) and show how the Minimax algorithm selects the best move for the maximizing player.

    View this question on its own page →
  18. Q5b. Compare probabilistic reasoning using Bayesian Networks with logical reasoning. In what situations is probabilistic reasoning more appropriate? Give examples.20247m

    Module 2: Knowledge Representation and Reasoning

    Compare probabilistic reasoning using Bayesian Networks with logical reasoning. In what situations is probabilistic reasoning more appropriate? Give examples.

    View this question on its own page →
  19. Q6a. What is partial-order planning? Explain with a block-world example, showing how actions can be arranged without a strict linear order.20247m

    Module 2: Knowledge Representation and Reasoning

    What is partial-order planning? Explain with a block-world example, showing how actions can be arranged without a strict linear order.

    View this question on its own page →
  20. Q6b. Explain resolution in First-Order Logic (FOL). Using a simple knowledge base (e.g., "All humans are mortal, Socrates is a human"), derive a conclusion using resolution refutation.20247m

    Module 2: Knowledge Representation and Reasoning

    Explain resolution in First-Order Logic (FOL). Using a simple knowledge base (e.g., "All humans are mortal, Socrates is a human"), derive a conclusion using resolution refutation.

    View this question on its own page →
  21. Q7a. Construct a simple decision tree from the given dataset: | Weather | Temp | Play? | |---|---|---| | Sunny | Hot | No | | Sunny | Cool | Yes | | Rainy | Cool | Yes | | Rainy | Hot | No | Explain your attribute selection criteria.20247m

    Module 3: Machine Learning Fundamentals

    Construct a simple decision tree from the given dataset:

    Weather Temp Play?
    Sunny Hot No
    Sunny Cool Yes
    Rainy Cool Yes
    Rainy Hot No

    Explain your attribute selection criteria.

    View this question on its own page →
  22. Q7b. How can semi-supervised learning be useful in real-world scenarios like email spam detection? Describe its working with an example involving labeled and unlabeled data.20247m

    Module 3: Machine Learning Fundamentals

    How can semi-supervised learning be useful in real-world scenarios like email spam detection? Describe its working with an example involving labeled and unlabeled data.

    View this question on its own page →
  23. Q8a. Discuss how backpropagation helps train a neural network. Use a small example with two layers and a loss function to explain weight adjustment.20247m

    Module 3: Machine Learning Fundamentals

    Discuss how backpropagation helps train a neural network. Use a small example with two layers and a loss function to explain weight adjustment.

    View this question on its own page →
  24. Q8b. What are the limitations of K-means clustering? Discuss two scenarios where K-means performs poorly and suggest possible improvements or alternatives.20247m

    Module 3: Machine Learning Fundamentals

    What are the limitations of K-means clustering? Discuss two scenarios where K-means performs poorly and suggest possible improvements or alternatives.

    View this question on its own page →
  25. Q9a. What is a Genetic Algorithm (GA)? Outline the basic steps of a GA such as selection, crossover, and mutation using an example of solving a simple optimization problem.20247m

    Module 4: Advanced AI Applications

    What is a Genetic Algorithm (GA)? Outline the basic steps of a GA such as selection, crossover, and mutation using an example of solving a simple optimization problem.

    View this question on its own page →
  26. Q9b. Explain the working of rule-based Expert Systems. Create a small rule base for diagnosing common cold vs. flu using IF-THEN rules.20247m

    Module 4: Advanced AI Applications

    Explain the working of rule-based Expert Systems. Create a small rule base for diagnosing common cold vs. flu using IF-THEN rules.

    View this question on its own page →