0.25
0.5
0.75
1.25
1.5
1.75
2
Variational Autoencoder and Extensions
Published on Sep 13, 201518495 Views
Related categories
Chapter list
Variational Autoencoder and Extensions 00:00
Outline00:53
Deep directed graphical models02:14
Latent variable generative model 03:13
Variational autoencoder (VAE) approach07:11
What VAE can do?08:54
The inference / learning challenge10:47
Variational Autoencoder (VAE)12:00
Reparametrization trick30:24
Training with backpropagation! 30:25
Relative performance of VAE32:48
Effect of KL term: component collapse33:48
VAE Inference model 45:56
Component collapse & decoder weights46:35
Semi-supervised Learning with Deep Generative Models - 153:10
Semi-supervised Learning with Deep Generative Models - 255:45
Semi-supervised Learning with Deep Generative Models - 456:00
Semi-supervised Learning with Deep Generative Models - 557:39
Semi-supervised MNIST classification results59:12
Semi-supervised Learning with Deep Generative Models - 301:01:36
Conditional generation using M201:02:13
A Recurrent Latent Variable Model for Sequential Data 01:03:40
VRNN: Model Structure - 101:03:43
VRNN: Model Structure - 201:05:31
VRNN: Prior on zt01:05:49
VRNN: Generation01:06:25
VRNN: Recurrence01:07:13
VRNN: Inference01:07:27
VRNN: Learning01:07:53
VRNN: Results01:09:15
VRNN: Speech synthesis01:10:25
VRNN: KL Divergence01:11:28
VRNN: Writing synthesis01:11:55
DRAW Deep Recurrent Attentive Writer 01:12:56
DRAW: Deep Recurrent Attentive Writer01:13:10
Variational Autoencoder Recap01:14:15
DRAW Model01:17:52
DRAW MNIST Generation01:17:56
DRAW Attention Mechanism - 101:18:37
DRAW Attention Mechanism - 201:20:38
DRAW: Cluttered MNIST Classification01:21:15
DRAW MNIST Generation with Attention - 101:22:39
DRAW MNIST Generation with Attention - 201:25:03
Recent Innovations in VAE Inference01:27:58
Inference in the VAE01:28:07
Variational Inference with Normalizing Flows01:28:41
Normalizing Flows - 101:28:46
Normalizing Flow - 201:29:35
Normalizing Flows for VAE posteriors - 101:29:44
Normalizing Flows for VAE posteriors - 201:30:04
Normalizing Flows for VAE posteriors - 301:30:44
Markov Chain Monte Carlo and Variational Inference: Bridging the Gap01:31:33
Variational and MCMC inference01:31:39
Hamiltonian (Hybrid) Monte Carlo - 101:31:39
Hamiltonian (Hybrid) Monte Carlo - 201:32:01
HMC for Deep Generative Models01:32:02
Hamiltonian Variational Inference (HVI) - 101:32:03
Hamiltonian Variational Inference (HVI) - 201:32:25
Hamiltonian Variational Inference (HVI) - 301:32:27
HVI: Generative model of MNIST01:32:32
The end01:32:43