FACT or Fiction: Can Truthful Mechanisms Eliminate Federated Free Riding?

TL;DR

This paper introduces FACT, a novel federated learning mechanism that prevents free-riding and false information submission by incentivizing truthful participation and outperforming training alone.

Contribution

FACT is the first federated mechanism to simultaneously eliminate free-riding, ensure truthfulness, and promote participation through a penalty system and competitive environment.

Findings

FACT prevents free-riding even with untruthful agents

Reduces agent loss by over 4 times

Outperforms training alone in empirical tests

Abstract

Standard federated learning (FL) approaches are vulnerable to the free-rider dilemma: participating agents can contribute little to nothing yet receive a well-trained aggregated model. While prior mechanisms attempt to solve the free-rider dilemma, none have addressed the issue of truthfulness. In practice, adversarial agents can provide false information to the server in order to cheat its way out of contributing to federated training. In an effort to make free-riding-averse federated mechanisms truthful, and consequently less prone to breaking down in practice, we propose FACT. FACT is the first federated mechanism that: (1) eliminates federated free riding by using a penalty system, (2) ensures agents provide truthful information by creating a competitive environment, and (3) encourages agent participation by offering better performance than training alone. Empirically, FACT avoids free-riding when agents are untruthful, and reduces agent loss by over 4x.