Questions:

1. explain SGD
2. explain SVD
   1. if A is a $m\ast n$ matrix, what's the shape of U, V
   2. do you know any properties of these matrix?
3. given a function graph with two local minimum, when GD (with fixed lr) fall into the far one?
4. which module (homework) you like the most?
5. Given a function, calculate the gradient.
6. What is normal distribution? (formula)
7. GD with backtracking
8. PCA
9. Sgd, standardisation, bias column, mse and bce, hw2 ex4
10. explain backtracking / armijo
11. explain convex function
12. explain the output of hw1 exercise 2
13. some clarification about HW3
14. gradient descent/ SGD / Adam with 1 data point, what changes
15. MAP
16. PCA, the maths behind it, the code analysis and explaining the PCA homework plots. Also, he asked me about clustering and its relation in PCA alongside its maths.
17. GD, SGD and Adam on a 1 sample batch. Eckart Young and PCA, and derive e^(log(x^2))
18. SVD, Matrix approximation, Young Theorem and 4.2
19. hw2 and 4, 2 in general, then explain SVD the formula, why do we do the centroid