FedMABA: Towards Fair Federated Learning through Multi-Armed Bandits Allocation
This addresses fairness issues in federated learning for privacy-sensitive applications, but it is incremental as it builds on existing FL methods with a novel allocation approach.
The paper tackles performance unfairness in federated learning due to statistical heterogeneity among clients, proposing FedMABA, a multi-armed bandit-based allocation algorithm that reduces performance disparities, with experiments showing enhanced fairness in Non-I.I.D. scenarios.
The increasing concern for data privacy has driven the rapid development of federated learning (FL), a privacy-preserving collaborative paradigm. However, the statistical heterogeneity among clients in FL results in inconsistent performance of the server model across various clients. Server model may show favoritism towards certain clients while performing poorly for others, heightening the challenge of fairness. In this paper, we reconsider the inconsistency in client performance distribution and introduce the concept of adversarial multi-armed bandit to optimize the proposed objective with explicit constraints on performance disparities. Practically, we propose a novel multi-armed bandit-based allocation FL algorithm (FedMABA) to mitigate performance unfairness among diverse clients with different data distributions. Extensive experiments, in different Non-I.I.D. scenarios, demonstrate the exceptional performance of FedMABA in enhancing fairness.