Byzantine Robustness and Partial Participation Can Be Achieved at Once: Just Clip Gradient Differences

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Malinovsky, Grigory, Richtárik, Peter, Horváth, Samuel, Gorbunov, Eduard
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913380895293440
author Malinovsky, Grigory
Richtárik, Peter
Horváth, Samuel
Gorbunov, Eduard
author_facet Malinovsky, Grigory
Richtárik, Peter
Horváth, Samuel
Gorbunov, Eduard
contents Distributed learning has emerged as a leading paradigm for training large machine learning models. However, in real-world scenarios, participants may be unreliable or malicious, posing a significant challenge to the integrity and accuracy of the trained models. Byzantine fault tolerance mechanisms have been proposed to address these issues, but they often assume full participation from all clients, which is not always practical due to the unavailability of some clients or communication constraints. In our work, we propose the first distributed method with client sampling and provable tolerance to Byzantine workers. The key idea behind the developed method is the use of gradient clipping to control stochastic gradient differences in recursive variance reduction. This allows us to bound the potential harm caused by Byzantine workers, even during iterations when all sampled clients are Byzantine. Furthermore, we incorporate communication compression into the method to enhance communication efficiency. Under general assumptions, we prove convergence rates for the proposed method that match the existing state-of-the-art (SOTA) theoretical results. We also propose a heuristic on adjusting any Byzantine-robust method to a partial participation scenario via clipping.
format Preprint
id arxiv_https___arxiv_org_abs_2311_14127
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Byzantine Robustness and Partial Participation Can Be Achieved at Once: Just Clip Gradient Differences
Malinovsky, Grigory
Richtárik, Peter
Horváth, Samuel
Gorbunov, Eduard
Machine Learning
Artificial Intelligence
Distributed, Parallel, and Cluster Computing
Optimization and Control
Distributed learning has emerged as a leading paradigm for training large machine learning models. However, in real-world scenarios, participants may be unreliable or malicious, posing a significant challenge to the integrity and accuracy of the trained models. Byzantine fault tolerance mechanisms have been proposed to address these issues, but they often assume full participation from all clients, which is not always practical due to the unavailability of some clients or communication constraints. In our work, we propose the first distributed method with client sampling and provable tolerance to Byzantine workers. The key idea behind the developed method is the use of gradient clipping to control stochastic gradient differences in recursive variance reduction. This allows us to bound the potential harm caused by Byzantine workers, even during iterations when all sampled clients are Byzantine. Furthermore, we incorporate communication compression into the method to enhance communication efficiency. Under general assumptions, we prove convergence rates for the proposed method that match the existing state-of-the-art (SOTA) theoretical results. We also propose a heuristic on adjusting any Byzantine-robust method to a partial participation scenario via clipping.
title Byzantine Robustness and Partial Participation Can Be Achieved at Once: Just Clip Gradient Differences
topic Machine Learning
Artificial Intelligence
Distributed, Parallel, and Cluster Computing
Optimization and Control
url https://arxiv.org/abs/2311.14127