← 02 Math for ML
Area 02 test · Foundations

Math for ML

20 questions · 70% to pass · no time limit. You see the right answers and why after you submit.

  1. 1What is [2, 0, 1] · [3, 5, −2]?
  2. 2A linear model predicts with…
  3. 3Shape of X @ W with X (64, 784) and W (784, 128)?
  4. 4In NumPy, A * B for two matrices is…
  5. 5Cosine similarity of two vectors pointing in exactly opposite directions?
  6. 6Why is cosine similarity preferred for comparing text embeddings?
  7. 7The derivative of f(w) = (w − 3)² at w = 5 is…
  8. 8Gradient descent updates a weight by…
  9. 9The loss turns to NaN after a few steps. Most likely cause?
  10. 10The chain rule says the derivative of f(g(x)) is…
  11. 11Backpropagation is…
  12. 12How can you check a hand-written gradient?
  13. 13A condition affects 1% of people. A test catches 90% of cases and flags 5% of healthy people. P(sick | positive) is closest to…
  14. 14Softmax of the scores [1, 1, 1] is…
  15. 15Cross-entropy loss for a correct class predicted with probability 0.01 is…
  16. 16The sigmoid's derivative σ(x)(1−σ(x)) is at most…
  17. 17A feature is heavily skewed by a few huge values. Best summary of its centre?
  18. 18Feature usage and retention have r = 0.7. What can you conclude?
  19. 19Model A: 0.84 accuracy, model B: 0.83, on 50 test examples. What should you do?
  20. 20In the lab, the analytic dL/db is about n times too large. What is missing?
0 of 20 answered