pith. sign in

arxiv: 1901.08508 · v2 · pith:QYKGJIL5new · submitted 2019-01-24 · 💻 cs.LG · cs.AI· stat.ML

Maximum Entropy Generators for Energy-Based Models

classification 💻 cs.LG cs.AIstat.ML
keywords gradientcompetitiveenergy-basedentropylog-likelihoodmaximummodelsproposed
0
0 comments X
read the original abstract

Maximum likelihood estimation of energy-based models is a challenging problem due to the intractability of the log-likelihood gradient. In this work, we propose learning both the energy function and an amortized approximate sampling mechanism using a neural generator network, which provides an efficient approximation of the log-likelihood gradient. The resulting objective requires maximizing entropy of the generated samples, which we perform using recently proposed nonparametric mutual information estimators. Finally, to stabilize the resulting adversarial game, we use a zero-centered gradient penalty derived as a necessary condition from the score matching literature. The proposed technique can generate sharp images with Inception and FID scores competitive with recent GAN techniques, does not suffer from mode collapse, and is competitive with state-of-the-art anomaly detection techniques.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Explaining the effects of non-convergent sampling in the training of Energy-Based Models

    cs.LG 2023-01 unverdicted novelty 7.0

    EBMs trained with non-persistent short runs reproduce empirical data statistics via a precise dynamical process, not the equilibrium measure.