@inproceedings{shi2022gradient,
author = {Shi, Jiaxin and Zhou, Yuhao and Hwang, Jessica and Titsias, Michalis and Mackey, Lester},
title = {Gradient Estimation with Discrete Stein Operators},
booktitle = {NeurIPS 2022},
year = {2022},
month = {November},
abstract = {Gradient estimation -- approximating the gradient of an expectation with respect to the parameters of a distribution -- is central to the solution of many machine learning problems. However, when the distribution is discrete, most common gradient estimators suffer from excessive variance. To improve the quality of gradient estimation, we introduce a variance reduction technique based on Stein operators for discrete distributions. We then use this technique to build flexible control variates for the REINFORCE leave-one-out estimator. Our control variates can be adapted online to minimize variance and do not require extra evaluations of the target function. In benchmark generative modeling tasks such as training binary variational autoencoders, our gradient estimator achieves substantially lower variance than state-of-the-art estimators with the same number of function evaluations.},
url = {http://approjects.co.za/?big=en-us/research/publication/gradient-estimation-with-discrete-stein-operators/},
}