Abstract:Adversarial learning methods have been proposed for a wide range of applications, but the training of adversarial models can be notoriously unstable. Effectively balancing the performance of the generator and discriminator is critical, since a discriminator that achieves very high accuracy will produce relatively uninformative gradients. In this work, we propose a simple and general technique to constrain information flow in the discriminator by means of an information bottleneck. By enforcing a constraint on the mutual information between the observations and the discriminator's internal representation, we can effectively modulate the discriminator's accuracy and maintain useful and informative gradients. We demonstrate that our proposed variational discriminator bottleneck (VDB) leads to significant improvements across three distinct application areas for adversarial learning algorithms. Our primary evaluation studies the applicability of the VDB to imitation learning of dynamic continuous control skills, such as running. We show that our method can learn such skills directly from \emph{raw} video demonstrations, substantially outperforming prior adversarial imitation learning methods. The VDB can also be combined with adversarial inverse reinforcement learning to learn parsimonious reward functions that can be transferred and re-optimized in new settings. Finally, we demonstrate that VDB can train GANs more effectively for image generation, improving upon a number of prior stabilization methods.

Variational Predictive Information Bottleneck

Tighter Bounds on the Information Bottleneck with Application to Deep Learning

On the Difference Between the Information Bottleneck and the Deep Information Bottleneck

Variational Information Bottleneck on Vector Quantized Autoencoders

Generalized Information Bottleneck for Gaussian Variables

Uncertainty in the Variational Information Bottleneck

Deep Variational Multivariate Information Bottleneck -- A Framework for Variational Losses

Relevant sparse codes with variational information bottleneck

A Variational Information Bottleneck Method for Network Intrusion Detection

Flexible Variational Information Bottleneck: Achieving Diverse Compression with a Single Training

Visualizing Information Bottleneck through Variational Inference

Variational Discriminator Bottleneck: Improving Imitation Learning, Inverse RL, and GANs by Constraining Information Flow

Generalized Variational Inference Via Optimal Transport

Information Bottleneck Revisited: Posterior Probability Perspective with Optimal Transport

Elastic Information Bottleneck

Variational Inference: A Review for Statisticians

On Variational Bounds of Mutual Information

Efficient and Provably Convergent Computation of Information Bottleneck: A Semi-Relaxed Approach

Caveats for information bottleneck in deterministic scenarios

Predictive variational inference: Learn the predictively optimal posterior distribution

Distributed Variational Information Bottleneck for IOT Environments