Power to the Clients: Federated Learning in a Dictatorship Setting

TL;DR

This paper introduces a new class of malicious clients in federated learning called dictators, capable of erasing other clients' contributions, analyzes their impact theoretically, and validates findings empirically on vision and NLP tasks.

Contribution

It defines and analyzes dictator clients in federated learning, proposing attack strategies and exploring complex multi-dictator scenarios with theoretical and empirical insights.

Findings

Dictator clients can fully erase other clients' contributions.

Multiple dictators can collaborate or betray, affecting convergence.

Empirical results confirm theoretical predictions on vision and NLP benchmarks.

Abstract

Federated learning (FL) has emerged as a promising paradigm for decentralized model training, enabling multiple clients to collaboratively learn a shared model without exchanging their local data. However, the decentralized nature of FL also introduces vulnerabilities, as malicious clients can compromise or manipulate the training process. In this work, we introduce dictator clients, a novel, well-defined, and analytically tractable class of malicious participants capable of entirely erasing the contributions of all other clients from the server model, while preserving their own. We propose concrete attack strategies that empower such clients and systematically analyze their effects on the learning process. Furthermore, we explore complex scenarios involving multiple dictator clients, including cases where they collaborate, act independently, or form an alliance in order to ultimately betray one another. For each of these settings, we provide a theoretical analysis of their impact on the global model's convergence. Our theoretical algorithms and findings about the complex scenarios including multiple dictator clients are further supported by empirical evaluations on both computer vision and natural language processing benchmarks.