Completed
Intro
Class Central Classrooms beta
YouTube playlists curated by Class Central.
Classroom Contents
Beyond Lazy Training for Over-parameterized Tensor Decomposition
Automatically move to the next video in the Classroom when playback concludes
- 1 Intro
- 2 Low rank models and implicit regularizati
- 3 Regimes of over-parametrization
- 4 Tensor (CP) decomposition
- 5 Why naïve algorithm fails
- 6 Why gradient descent?
- 7 Two-Layer Neural Network
- 8 Form of the objective
- 9 Difficulties of analyzing gradient descent
- 10 Lazy training fails
- 11 O is a high order saddle point
- 12 There are local minima away from 0
- 13 Our (high level) algorithm
- 14 Proof ideas
- 15 Escaping local minima by random correla
- 16 Amplify initial correlation by tensor power man
- 17 Conclusions and Open Problems
