Discriminator Contrastive Divergence: Semi-Amortized Generative Modeling by Exploring Energy of the Discriminator

Abstract

Generative Adversarial Networks (GANs) have shown great promise in modeling high dimensional data. The learning objective of GANs usually minimizes some measure discrepancy, \textit{e.g.}, $f$-divergence~($f$-GANs) or Integral Probability Metric~(Wasserstein GANs). With $f$-divergence as the objective function, the discriminator essentially estimates the density ratio, and the estimated ratio proves useful in further improving the sample quality of the generator. However, how to leverage the information contained in the discriminator of Wasserstein GANs (WGAN) is less explored. In this paper, we introduce the Discriminator Contrastive Divergence, which is well motivated by the property of WGAN's discriminator and the relationship between WGAN and energy-based model. Compared to standard GANs, where the generator is directly utilized to obtain new samples, our method proposes a semi-amortized generation procedure where the samples are produced with the generator's output as an initial state. Then several steps of Langevin dynamics are conducted using the gradient of the discriminator. We demonstrate the benefits of significant improved generation on both synthetic data and several real-world image generation benchmarks.

17 pages, 9 figures, pre-submmited to cvpr2019

References in corpus (23)

• Adam: A Method for Stochastic Optimization
• GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium
• Large Scale GAN Training for High Fidelity Natural Image Synthesis
• Spectral Normalization for Generative Adversarial Networks
• Conditional Image Generation with PixelCNN Decoders
• Progressive Growing of GANs for Improved Quality, Stability, and Variation
• Energy-based Generative Adversarial Network
• Density estimation using Real NVP
• Estimating divergence functionals and the likelihood ratio by convex risk minimization
• A note on the evaluation of generative models
• A Connection between Generative Adversarial Networks, Inverse Reinforcement Learning, and Energy-Based Models
• Boundary-Seeking Generative Adversarial Networks
• Implicit Generation and Generalization in Energy-Based Models
• Your Classifier is Secretly an Energy Based Model and You Should Treat it Like One
• Deep Directed Generative Models with Energy-Based Probability Estimation
• Bias Correction of Learned Generative Models using Likelihood-Free Importance Weighting
• Maximum Entropy Generators for Energy-Based Models
• Generative Adversarial Nets from a Density Ratio Estimation Perspective
• Lipschitz Generative Adversarial Nets
• Autoregressive Quantile Networks for Generative Modeling
• Approximate Inference with Amortised MCMC
• Metropolis-Hastings Generative Adversarial Networks
• Learning Implicit Generative Models with the Method of Learned Moments

Cited by in corpus (5)

• Deep Generative Modelling: A Comparative Review of VAEs, GANs, Normalizing Flows, Energy-Based and Autoregressive Models
• Learning Neural Generative Dynamics for Molecular Conformation Generation
• VAEBM: A Symbiosis between Variational Autoencoders and Energy-based Models
• EBMs Trained with Maximum Likelihood are Generator Models Trained with a Self-adverserial Loss
• Reparameterized Sampling for Generative Adversarial Networks