ByRDiE: Byzantine-resilient distributed coordinate descent for decentralized learning

TL;DR

This paper introduces ByRDiE, a new distributed coordinate descent algorithm designed to be resilient against Byzantine failures, enabling reliable high-dimensional decentralized learning despite malicious or faulty network nodes.

Contribution

The paper presents the first practical Byzantine-resilient distributed coordinate descent algorithm for high-dimensional decentralized learning, with theoretical guarantees and empirical validation.

Findings

ByRDiE achieves Byzantine resilience in distributed learning.

The algorithm performs well in convex and nonconvex settings.

Numerical experiments confirm its effectiveness in high-dimensional scenarios.

Abstract

Distributed machine learning algorithms enable learning of models from datasets that are distributed over a network without gathering the data at a centralized location. While efficient distributed algorithms have been developed under the assumption of faultless networks, failures that can render these algorithms nonfunctional occur frequently in the real world. This paper focuses on the problem of Byzantine failures, which are the hardest to safeguard against in distributed algorithms. While Byzantine fault tolerance has a rich history, existing work does not translate into efficient and practical algorithms for high-dimensional learning in fully distributed (also known as decentralized) settings. In this paper, an algorithm termed Byzantine-resilient distributed coordinate descent (ByRDiE) is developed and analyzed that enables distributed learning in the presence of Byzantine failures. Theoretical analysis (convex settings) and numerical experiments (convex and nonconvex settings) highlight its usefulness for high-dimensional distributed learning in the presence of Byzantine failures.