Question 1

How does the Dueling DQN architecture enhance the learning of value functions?

Options :

A : By incorporating a second experience replay buffer for improved stability

B : By combining Q-values with advantage values to better separate value and advantage estimates

C : By adding a recurrent layer to capture temporal dependencies

D : By using a separate network to compute the discount factor for future rewards

Answer: B

Question 2

What is a significant ethical challenge associated with the deployment of White Box XAI models?

Options :

A : Risk of revealing sensitive data through transparency

B : Difficulty in achieving model accuracy

C : Potential over-reliance on human interpretations

D : Complexity in maintaining model interpretability

Answer: A

Question 3

In dynamic programming, what does the term "policy evaluation" refer to?

Options :

A : Comparing different policies for effectiveness

B : Updating the policy based on new data

C : Selecting the best action in each state

D : Calculating the value function for a given policy

Answer: D

Question 4

What is the key advantage of the "Thompson Sampling" method over the "Upper Confidence Bound" (UCB) method in Multi-Armed Bandits?

Options :

A : Thompson Sampling is simpler to implement

B : Thompson Sampling performs better in environments with large action spaces

C : Thompson Sampling naturally handles uncertainty in reward estimates

D : Thompson Sampling does not require parameter tuning

Answer: C

Question 5

What is a common strategy to handle a model that is suffering from underfitting?

Options :

A : Increase model complexity by adding more layers or neurons

B : Simplify the model architecture

C : Apply data augmentation

D : Reduce the training dataset size

Answer: A