Training Discrete Deep Generative Models via Gapped Straight-Through Estimator
Abstract
While deep generative models have succeeded in image processing, natural language processing, and reinforcement learning, training that involves discrete random variables remains challenging due to the high variance of its gradient estimation process. Monte Carlo is a common solution used in most variance reduction approaches. However, this involves time-consuming resampling and multiple function evaluations. We propose a Gapped Straight-Through (GST) estimator to reduce the variance without incurring resampling overhead. This estimator is inspired by the essential properties of Straight-Through Gumbel-Softmax. We determine these properties and show via an ablation study that they are essential. Experiments demonstrate that the proposed GST estimator enjoys better performance compared to strong baselines on two discrete deep generative modeling tasks, MNIST-VAE and ListOps.
Cite
@article{arxiv.2206.07235,
title = {Training Discrete Deep Generative Models via Gapped Straight-Through Estimator},
author = {Ting-Han Fan and Ta-Chung Chi and Alexander I. Rudnicky and Peter J. Ramadge},
journal= {arXiv preprint arXiv:2206.07235},
year = {2022}
}
Comments
Accepted at the International Conference on Machine Learning (ICML) 2022. The first two authors contributed equally