Picture for Ben Poole

Ben Poole

VIB is Half Bayes

Add code
Nov 17, 2020
Figure 1 for VIB is Half Bayes
Figure 2 for VIB is Half Bayes
Figure 3 for VIB is Half Bayes
Viaarxiv icon

Non-saturating GAN training as divergence minimization

Add code
Oct 15, 2020
Figure 1 for Non-saturating GAN training as divergence minimization
Figure 2 for Non-saturating GAN training as divergence minimization
Figure 3 for Non-saturating GAN training as divergence minimization
Figure 4 for Non-saturating GAN training as divergence minimization
Viaarxiv icon

Tasks, stability, architecture, and compute: Training more effective learned optimizers, and using them to train themselves

Add code
Sep 23, 2020
Figure 1 for Tasks, stability, architecture, and compute: Training more effective learned optimizers, and using them to train themselves
Figure 2 for Tasks, stability, architecture, and compute: Training more effective learned optimizers, and using them to train themselves
Figure 3 for Tasks, stability, architecture, and compute: Training more effective learned optimizers, and using them to train themselves
Figure 4 for Tasks, stability, architecture, and compute: Training more effective learned optimizers, and using them to train themselves
Viaarxiv icon

What makes for good views for contrastive learning

Add code
May 20, 2020
Figure 1 for What makes for good views for contrastive learning
Figure 2 for What makes for good views for contrastive learning
Figure 3 for What makes for good views for contrastive learning
Figure 4 for What makes for good views for contrastive learning
Viaarxiv icon

Using a thousand optimization tasks to learn hyperparameter search strategies

Add code
Mar 11, 2020
Figure 1 for Using a thousand optimization tasks to learn hyperparameter search strategies
Figure 2 for Using a thousand optimization tasks to learn hyperparameter search strategies
Figure 3 for Using a thousand optimization tasks to learn hyperparameter search strategies
Figure 4 for Using a thousand optimization tasks to learn hyperparameter search strategies
Viaarxiv icon

Regularized Autoencoders via Relaxed Injective Probability Flow

Add code
Feb 20, 2020
Figure 1 for Regularized Autoencoders via Relaxed Injective Probability Flow
Figure 2 for Regularized Autoencoders via Relaxed Injective Probability Flow
Figure 3 for Regularized Autoencoders via Relaxed Injective Probability Flow
Figure 4 for Regularized Autoencoders via Relaxed Injective Probability Flow
Viaarxiv icon

On Implicit Regularization in $β$-VAEs

Add code
Feb 09, 2020
Figure 1 for On Implicit Regularization in $β$-VAEs
Figure 2 for On Implicit Regularization in $β$-VAEs
Figure 3 for On Implicit Regularization in $β$-VAEs
Figure 4 for On Implicit Regularization in $β$-VAEs
Viaarxiv icon

Weakly-Supervised Disentanglement Without Compromises

Add code
Feb 07, 2020
Figure 1 for Weakly-Supervised Disentanglement Without Compromises
Figure 2 for Weakly-Supervised Disentanglement Without Compromises
Figure 3 for Weakly-Supervised Disentanglement Without Compromises
Figure 4 for Weakly-Supervised Disentanglement Without Compromises
Viaarxiv icon

Weakly Supervised Disentanglement with Guarantees

Add code
Oct 22, 2019
Figure 1 for Weakly Supervised Disentanglement with Guarantees
Figure 2 for Weakly Supervised Disentanglement with Guarantees
Figure 3 for Weakly Supervised Disentanglement with Guarantees
Figure 4 for Weakly Supervised Disentanglement with Guarantees
Viaarxiv icon

On Predictive Information Sub-optimality of RNNs

Add code
Oct 21, 2019
Figure 1 for On Predictive Information Sub-optimality of RNNs
Figure 2 for On Predictive Information Sub-optimality of RNNs
Figure 3 for On Predictive Information Sub-optimality of RNNs
Figure 4 for On Predictive Information Sub-optimality of RNNs
Viaarxiv icon