Deep Learning Advanced Quiz 4

1. Which method is commonly used to visualize high-dimensional embeddings?

2. What is the main purpose of gradient clipping?

3. Which architecture is used in Deep Q-Networks for reinforcement learning?

4. What is the main idea of the Transformer’s multi-head attention?

5. Which technique is used to interpret CNN feature maps?

6. What is the main difference between batch normalization and layer normalization?

7. Which method is commonly used for self-supervised learning in NLP?

8. What is a main challenge when training very deep neural networks?

9. Which regularization method randomly mixes or swaps parts of inputs and labels?

10. What is the main purpose of positional encoding in Transformers?