Completed
Empirical Observations on Training Loss
Class Central Classrooms beta
YouTube videos curated by Class Central.
Classroom Contents
On the Connection Between Neural Networks and Kernels: A Modern Perspective - Simon Du
Automatically move to the next video in the Classroom when playback concludes
- 1 Intro
- 2 Two Fundamental Questions
- 3 Empirical Observations on Training Loss
- 4 Over-parameterization
- 5 Empirical Observations on Generalization
- 6 Example: Two-layer NN
- 7 Trajectory-based Analysis
- 8 The Trajectory of Predictions (Cont'd)
- 9 Kernel Matrix at the Beginning
- 10 Kernel Matrix During Training
- 11 Main Theory
- 12 Zero Training Error
- 13 Empirical Results on Generalization
- 14 Convolutional Neural Tangent Kernel
- 15 CNTK on CIFAR 10
- 16 Understanding Global Average Pooling
- 17 Local Average Pooling
- 18 UCI Experiment Setup
- 19 UCI Results
- 20 Few-shot Learning Setup
- 21 Few-shot Learning Results
- 22 Graph NTK for Graph Classification
- 23 Summary
- 24 References
