Implicit Generation and Modeling with Energy Based Models

NeurIPS 2019  ·  Yilun Du, Igor Mordatch ·

Energy based models (EBMs) are appealing due to their generality and simplicity in likelihood modeling, but have been traditionally difficult to train. We present techniques to scale MCMC based EBM training on continuous neural networks, and we show its success on the high-dimensional data domains of ImageNet32x32, ImageNet128x128, CIFAR-10, and robotic hand trajectories, achieving better samples than other likelihood models and nearing the performance of contemporary GAN approaches, while covering all modes of the data. We highlight some unique capabilities of implicit generation such as compositionality and corrupt image reconstruction and inpainting. Finally, we show that EBMs are useful models across a wide variety of tasks, achieving state-of-the-art out-of-distribution classification, adversarially robust classification, state-of-the-art continual online class learning, and coherent long term predicted trajectory rollouts.

PDF Abstract

Datasets


Results from the Paper


Results from Other Papers


Task Dataset Model Metric Name Metric Value Rank Source Paper Compare
Image Generation Stacked MNIST EBM on x FID 48.89 # 3
Inception score 5.85 # 2

Methods