Optimization Foundations

Derivatives, gradient descent and SGD, convexity, and positive semidefinite matrices.

0 / 8 solved
0/8 solved
Question Tags Difficulty Success
1. The Sharp Coordinate Sets the Step
gradient-descentlearning-rate +2
Medium —
2. NaN On Step 300
debuggingnumerical-stability +2
Medium —
3. Two Steps Downhill
gradient-descentoptimization +2
Easy 72%
4. Counting Work Not Updates
sgdminibatch +2
Medium —
5. Too Hot, Too Cold
learning rategradient descent
Easy —
6. Escaping Along the Negative Curvature
saddle-pointhessian +2
Medium —
7. No Valleys, Just One Basin
convexitypsd matrices +1
Medium —
8. Velocity Settles At Ten Times
momentumsgd +2
Medium —

Up next

8 of 8 questions here are still unsolved.