First Order Methods with Markovian Noise: from Acceleration to Variational Inequalities

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Beznosikov, Aleksandr, Samsonov, Sergey, Sheshukova, Marina, Gasnikov, Alexander, Naumov, Alexey, Moulines, Eric
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910391480614912
author Beznosikov, Aleksandr
Samsonov, Sergey
Sheshukova, Marina
Gasnikov, Alexander
Naumov, Alexey
Moulines, Eric
author_facet Beznosikov, Aleksandr
Samsonov, Sergey
Sheshukova, Marina
Gasnikov, Alexander
Naumov, Alexey
Moulines, Eric
contents This paper delves into stochastic optimization problems that involve Markovian noise. We present a unified approach for the theoretical analysis of first-order gradient methods for stochastic optimization and variational inequalities. Our approach covers scenarios for both non-convex and strongly convex minimization problems. To achieve an optimal (linear) dependence on the mixing time of the underlying noise sequence, we use the randomized batching scheme, which is based on the multilevel Monte Carlo method. Moreover, our technique allows us to eliminate the limiting assumptions of previous research on Markov noise, such as the need for a bounded domain and uniformly bounded stochastic gradients. Our extension to variational inequalities under Markovian noise is original. Additionally, we provide lower bounds that match the oracle complexity of our method in the case of strongly convex optimization problems.
format Preprint
id arxiv_https___arxiv_org_abs_2305_15938
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle First Order Methods with Markovian Noise: from Acceleration to Variational Inequalities
Beznosikov, Aleksandr
Samsonov, Sergey
Sheshukova, Marina
Gasnikov, Alexander
Naumov, Alexey
Moulines, Eric
Optimization and Control
Machine Learning
This paper delves into stochastic optimization problems that involve Markovian noise. We present a unified approach for the theoretical analysis of first-order gradient methods for stochastic optimization and variational inequalities. Our approach covers scenarios for both non-convex and strongly convex minimization problems. To achieve an optimal (linear) dependence on the mixing time of the underlying noise sequence, we use the randomized batching scheme, which is based on the multilevel Monte Carlo method. Moreover, our technique allows us to eliminate the limiting assumptions of previous research on Markov noise, such as the need for a bounded domain and uniformly bounded stochastic gradients. Our extension to variational inequalities under Markovian noise is original. Additionally, we provide lower bounds that match the oracle complexity of our method in the case of strongly convex optimization problems.
title First Order Methods with Markovian Noise: from Acceleration to Variational Inequalities
topic Optimization and Control
Machine Learning
url https://arxiv.org/abs/2305.15938