Deep Directed Generative Models with Energy-Based Probability Estimation

TL;DR

This paper introduces a deep directed generative model framework that uses energy-based probability estimation, avoiding complex MCMC sampling by training a generator and an energy function simultaneously, inspired by GANs.

Contribution

It proposes a novel training approach for energy-based models using deep neural networks without relying on Markov chain Monte Carlo sampling.

Findings

Effective approximation of energy functions with deep neural networks

Avoidance of intractable sums in energy-based models

Successful training of generator and energy function simultaneously

Abstract

Training energy-based probabilistic models is confronted with apparently intractable sums, whose Monte Carlo estimation requires sampling from the estimated probability distribution in the inner loop of training. This can be approximately achieved by Markov chain Monte Carlo methods, but may still face a formidable obstacle that is the difficulty of mixing between modes with sharp concentrations of probability. Whereas an MCMC process is usually derived from a given energy function based on mathematical considerations and requires an arbitrarily long time to obtain good and varied samples, we propose to train a deep directed generative model (not a Markov chain) so that its sampling distribution approximately matches the energy function that is being trained. Inspired by generative adversarial networks, the proposed framework involves training of two models that represent dual views of the estimated probability distribution: the energy function (mapping an input configuration to a scalar energy value) and the generator (mapping a noise vector to a generated configuration), both represented by deep neural networks.