Statistical inference with stochastic gradient algorithms

Files
2207.12395.pdf(1.38 MB)
First author draft
Date
2022-11-14
Authors
Negrea, Jeffrey
Yang, Jun
Feng, Haoyue
Roy, Daniel
Huggins, Jonathan
Version
First author draft
OA Version
Citation
J. Negrea, J. Yang, H. Feng, D. Roy, J. Huggins. 2022. "Statistical Inference with Stochastic Gradient Algorithms" https://doi.org/10.48550/arXiv.2207.12395
Abstract
The tuning of stochastic gradient algorithms (SGAs) for optimization and sampling is often based on heuristics and trial-and-error rather than generalizable theory. We address this theory–practice gap by characterizing the large-sample statistical asymptotics of SGAs via a joint step-size–sample-size scaling limit. We show that iterate averaging with a large fixed step size is robust to the choice of tuning parameters and asymptotically has covariance proportional to that of the MLE sampling distribution. We also prove a Bernstein–von Mises-like theorem to guide tuning, including for generalized posteriors that are robust to model misspecification. Numerical experiments validate our results and recommendations in realistic finite-sample regimes. Our work lays the foundation for a systematic analysis of other stochastic gradient Markov chain Monte Carlo algorithms for a wide range of models.
Description
License
This article is distributed under the terms of the Creative Commons Attribution 4.0 International (CC BY 4.0).