Paper 2 CS · UNIT 10 · MOCK TEST

Artificial Intelligence (AI) — Mock Test 3

25 questions · about 30 minutes. Press End Mock any time to see your score and the correct answers.

0 / 25 answered 00:00

Q1. What is a perceptron in an ANN?

AA double-layer auto-associative neural network BA neural network that contains feedback CAn auto-associative neural network DA single-layer feed-forward neural network with pre-processing

Q2. Arrange the following examples of AI in order of increasing complexity:
A. Spam e-mail detection using rule-based systems
B. Handwritten digit recognition using shallow neural networks
C. Image classification using convolutional neural networks
D. Autonomous driving using reinforcement-learning algorithms
Choose the correct answer from the options given below:

AA, B, C, D BB, A, C, D CA, B, D, C DB, C, A, D

Q3. For the following parents (crossover point after the 5th bit):
Figure (1): Parent₁ = 10110 | 010, Parent₂ = 10101 | 111
Figure (2): Parent₁ = 10110 | 111, Parent₂ = 10101 | 010
Which is correct?

AFigure (2) is a two-point crossover of figure (1) BFigure (2) is a one-point crossover of figure (1) CFigure (2) is an N-point crossover of figure (1) DFigure (2) is a uniform crossover of figure (1)

Q4. For the fuzzy set A = {(x1, 0.3), (x2, 0.7), (x3, 0.4)}, the complement of A would be:

A{(x1, 0.4), (x2, 0.3), (x3, 0.1)} B{(x1, 0.3), (x2, 0.7), (x3, 0.4)} C{(x1, 0.7), (x2, 0.3), (x3, 0.6)} D{(x1, 0.21), (x2, 0.28), (x3, 0.12)}

Q5. Backpropagation is a learning technique that adjusts weights in a neural network by propagating weight changes:

AForward from source to hidden nodes BBackward from sink to source CForward from source to sink DBackward from sink to hidden nodes

Q6. The cardinality of a (finite-universe) fuzzy set is:

A0 Bfinite Cinfinite Dnot known

Q7. The parameter 'actual count' is used in a Genetic Algorithm (GA) for:

ACrossover BMutation CSelecting population DEncoding the GA

Q8. Which of the following knowledge-representation structures are the most purely syntactic?
A. Conceptual dependency
B. Predicate logic
C. Slot-and-filler
D. Non-monotonic systems
Choose the correct answer from the options given below:

AA and B Only BB and C Only CB and D Only DC and D Only

Q9. In a Genetic Algorithm's crossover operation, a mask is used in:

AThree-parent crossover BTwo-parent crossover CUniform crossover DN-point crossover

Q10. For the fuzzy sets A = {(x1, 0.3), (x2, 0.7), (x3, 0.4)} and B = {(x1, 0.5), (x2, 0.2), (x3, 0.5)}, A ∩ B would be:

A{(x1, 0.3), (x2, 0.7), (x3, 0.4)} B{(x1, 0.5), (x2, 0.2), (x3, 0.5)} C{(x1, 0.5), (x2, 0.7), (x3, 0.5)} D{(x1, 0.3), (x2, 0.2), (x3, 0.4)}

Q11. Which of the following genetic-algorithm operations is computationally the most expensive?

AConvergence testing BInitial population creation CSelection of sub-population for mating DReproduction to produce the next generation

Q12. If we convert a decision tree into a set of logical rules, then:

AThe internal nodes in a branch are connected by AND and branches by AND BThe internal nodes in a branch are connected by OR and branches by OR CThe internal nodes in a branch are connected by AND and branches by OR DThe internal nodes in a branch are connected by OR and branches by AND

Q13. Which search technique is complete and optimal when h(n) is consistent?

ABest-first search BDepth-first search CA* search DBreadth-first search

Q14. The process of capturing an inference process as a single inference rule is called:

APonens BClauses CGeneralized modus ponens DVariables

Q15. Fuzzy logic is a form of ____

ATwo-valued logic BCrisp set logic CBinary set logic DMany-valued logic

Q16. The 4Ws problem canvas helps in identifying the key elements related to the problem. The 4Ws problem canvas is a part of ____

AProblem scoping BData acquisition CData modelling DProblem evaluation

Q17. Consider the following features of neural networks:
A. It is essentially a machine-learning algorithm.
B. It is useful for problems where the data set is very large.
C. It can extract features without input from the programmer.
D. It is a system modelled on the human brain and nervous system.
Choose the most appropriate answer from the options given below:

AA Only BA & B Only CA, B & C Only DA, B, C & D

Q18. Consider the following drawbacks of the rule-based approach:
A. The learning is static.
B. Changes made to the original data are not considered.
C. Once trained, the model cannot improve on the basis of feedback.

AAll the above statements are correct BStatements A & B only are not correct CStatements B & C only are not correct DAll the above statements are not correct

Q19. Match List I with List II (I. Unsupervised learning, II. Supervised learning): A. Regression, B. KNN, C. Clustering, D. Decision Tree

AA-I, B-II, C-I, D-II BA-II, B-II, C-I, D-II CA-I, B-II, C-II, D-I DA-II, B-II, C-II, D-I

Q20. Consider the weighted directed graph with edges A→C (2), A→F (7), C→B (10), C→F (3), B→D (1), B→Z (15), F→D (1), F→E (6), D→Z (7), D→E (2), E→Z (4). In what order are the nodes traversed to find the shortest path from A to Z?

AA → C → F → D → E BA → C → E → B CA → C → F → E → B DA → C → D → F

Q21. Which of the following describes the flow through the layers of a neural network?
A. Information → Input layer → Hidden layer
B. Bias → Output layer → Hidden layer
C. Hidden layer → Algorithm → Output layer
D. Input layer → Output layer → Algorithms

AOnly sequence A is correct. BOnly sequence B is correct. COnly sequence C is correct. DOnly sequence D is correct.

Q22. Given below are two statements: one is labelled as Assertion A and the other as Reason R.
A: Artificial neural networks are the backbone of deep-learning algorithms.
R: Back propagation is one of the techniques used to train an ANN.

ABoth A and R are correct and R is the correct explanation of A. BBoth A and R are correct but R is NOT the correct explanation of A. CA is correct but R is not correct. DA is not correct but R is correct.

Q23. Overfitting is expected when we observe that

AWith training iterations, errors on the training set as well as the test set decrease BWith training iterations, error on the training set decreases but on the test set increases CWith training iterations, errors on the training set as well as the test set increase DWith training iterations, training as well as test error remain constant

Q24. The A* algorithm is optimal when

Ait always finds the solution with the lowest total cost if the heuristic h is admissible Bit always finds the solution with the highest total cost if the heuristic h is admissible Cit finds the solution with the lowest total cost if the heuristic h is not admissible Dit always finds the solution with the highest total cost if the heuristic h is not admissible

Q25. Which artificial intelligence technique enables computers to understand the associations and relationships between objects and events?

AHeuristic Processing BCognitive Science CRelative Symbolism DPattern Matching