Translator Disclaimer
May 2022 Oracle lower bounds for stochastic gradient sampling algorithms
Niladri S. Chatterji, Peter L. Bartlett, Philip M. Long
Author Affiliations +
Bernoulli 28(2): 1074-1092 (May 2022). DOI: 10.3150/21-BEJ1377

Abstract

We consider the problem of sampling from a strongly log-concave density in Rd, and prove an information theoretic lower bound on the number of stochastic gradient queries of the log density needed. Several popular sampling algorithms (including many Markov chain Monte Carlo methods) operate by using stochastic gradients of the log density to generate a sample; our results establish an information theoretic limit for all these algorithms.

We show that for every algorithm, there exists a well-conditioned strongly log-concave target density for which the distribution of points generated by the algorithm would be at least ε away from the target in total variation distance if the number of gradient queries is less than Ω(σ2d/ε2), where σ2d is the variance of the stochastic gradient. Our lower bound follows by combining the ideas of Le Cam deficiency routinely used in the comparison of statistical experiments along with standard information theoretic tools used in lower bounding Bayes risk functions. To the best of our knowledge our results provide the first nontrivial dimension-dependent lower bound for this problem.

Funding Statement

We gratefully acknowledge the support of the NSF through grants IIS-1619362 and IIS-1909365. Part of this work was completed while NC was interning at Google.

Acknowledgements

We are grateful to Aditya Guntuboyina for pointing us towards the literature on Le Cam deficiency. We would also like to thank Jelena Diakonikolas, Sébastien Gerchinovitz, Michael Jordan, Aldo Pacchiano, Aaditya Ramdas and Morris Yau for many helpful conversations. We thank Kush Bhatia for helpful comments that improved the presentation of the results.

Citation

Download Citation

Niladri S. Chatterji. Peter L. Bartlett. Philip M. Long. "Oracle lower bounds for stochastic gradient sampling algorithms." Bernoulli 28 (2) 1074 - 1092, May 2022. https://doi.org/10.3150/21-BEJ1377

Information

Received: 1 October 2020; Revised: 1 March 2021; Published: May 2022
First available in Project Euclid: 3 March 2022

Digital Object Identifier: 10.3150/21-BEJ1377

Keywords: information theoretic lower bounds , Markov chain Monte Carlo , Sampling lower bounds , stochastic gradient Monte Carlo

Rights: Copyright © 2022 ISI/BS

JOURNAL ARTICLE
19 PAGES

This article is only available to subscribers.
It is not available for individual sale.
+ SAVE TO MY LIBRARY

SHARE
Vol.28 • No. 2 • May 2022
Back to Top