Meta Stackelberg Game: Robust Federated Learning against Adaptive and Mixed Poisoning Attacks

Henger Li, Tao Li, Yunian Pan, Tianyi Xu, Quanyan Zhu, Zizhan Zheng

Keywords: Federated Learning, Game Theory, Reinforcement Learning, Robust Machine Learning

Abstract: Recent research has uncovered that federated learning (FL) systems are vulnerable to various security threats. Although various defense mechanisms have been proposed, they are typically non-adaptive and tailored to specific types of attacks, leaving them insufficient in the face of adaptive or mixed attacks. In this work, we formulate adversarial federated learning as a Bayesian Stackelberg Markov game (BSMG) to tackle poisoning attacks of unknown/uncertain types. We further develop an efficient meta-learning approach to solve the game, which provides a robust and adaptive FL defense. Theoretically, we show that our algorithm provably converges to the first-order $\varepsilon$-equilibrium point in $O(\varepsilon^{-2})$ gradient iterations with $O(\varepsilon^{-4})$ samples per iteration. Empirical results show that our meta-Stackelberg framework obtains superb performance against strong model poisoning and backdoor attacks with unknown/uncertain types.

Primary Area: Safety in machine learning

Submission Number: 13590