CS 540 Spring 2024 Final

Content locked

The catalog is public. Reading this content requires approved access.

Question 1 (2D convolution)

Question 2 (pooling types)

Question 3 (conv output size)

Question 4 (CNN vs. MLP — NOT a reason)

Question 5 (vanishing gradient fix)

Question 6 (deep nets — NOT a reason for worse performance)

Question 7 (bad data augmentation)

Question 8 (regularization — FALSE)

Question 9 (residual networks — TRUE)

Question 10 (BFS data structure)

Question 11 (infinite-depth graph)

Question 12 (uninformed search behavior)

Question 13 (leaky ReLU)

Question 14 (perceptron + ReLU)

Question 15 (single-layer net, matrix form)

Question 16 (admissible heuristic on \mathbb{Z})

Question 17 (dominant strategies)

Question 18 (Nash equilibrium)

Question 19 (minimax facts)

Question 20 (counting minimax leaves)

Question 21 (RL domains)

Question 22 (discount factor)

Question 23 (Markov property)

Question 24 (ε-greedy Q-learning — NOT true)

Question 25 (one value-iteration step)