Completed
Intro
Class Central Classrooms beta
YouTube playlists curated by Class Central.
Classroom Contents
Deep Neural Networks via Monotone Operators
Automatically move to the next video in the Classroom when playback concludes
- 1 Intro
- 2 The deep learning revolution recent examp
- 3 Deep Learning The story we all tell: deep learning algorithms build hierarchical models of input date, where the earlier layers create simple features and layer layers create high- level abstractions…
- 4 This talk
- 5 Outline
- 6 From deep networks to DEQs
- 7 Long history of related work
- 8 Implementing DEQS
- 9 The DEQ forward pass
- 10 How to train your DEQ
- 11 How to train your DED Compute gradients analytically via implicit function theorem
- 12 More information on implicit layers
- 13 Language modeling: WikiText-103
- 14 Multiscale deep equilibrium models
- 15 ImageNet Top-1 Accuracy
- 16 Citiscapes mlou
- 17 Visualization of Segmentation
- 18 Theoretical/algorithmic challenges for DE
- 19 Key result
- 20 Proof sketch for simpler case
- 21 Monotone operator equilibrium network
- 22 Initial study: CIFAR10
- 23 Additional points on monotone DEOS
- 24 Final thoughts
