sp26-cs639 cs639
Practice Midterm 4 Practice Midterm 4
Content locked
The catalog is public. Reading this content requires approved access.
Sign in to read1. [ML Foundations: Hypothesis Classes] Capacity and Generalization
2. [ML Foundations: Logistic Regression] Sigmoid Boundary
3. [Deep Learning: CNN] Receptive Field Growth
4. [Deep Learning: RNN] Weight Sharing Across Time
5. [Deep Learning: GNN] Multi-Hop Information
6. [Self-Supervised Learning: Pretext Tasks] Masking Strategy Design
7. [Self-Supervised Learning: Contrastive] Negative Sampling Pitfall
8. [Scaling Laws] Validation Set Mismatch
9. [Scaling Laws] Iso-FLOP Interpretation
10. [Transformers: Attention] Permutation Equivariance
11. [Transformers: Architecture] Residual Connection Purpose
12. [Transformers: Multi-Head] Parameter Count
13. [BERT: Training] Masking Rate Rationale
14. [BERT vs GPT] Architectural Asymmetry
15. [GPT: Architecture] Decoder Block Components
16. [Llama: Architecture] Grouped Query Attention
17. [Attention Variants: Sliding Window] Information Propagation
18. [FlashAttention] Computational Trade-off
19. [SSMs: Mamba] Selectivity Trade-off
20. [Multimodal: ViT] Patch Embedding Computation
21. [Multimodal: CLIP vs VisualBERT] Fusion Timing
22. [Multimodal: LLaVA] Architecture Components
23. [Prompting: ICL] Example Selection Impact
24. [Prompting: Chain-of-Thought] When CoT Fails
25. [Prompting: Tree of Thoughts] Search vs Sampling