Toward Personalized Federated Learning via Overlapping Coalition Formation Game

To tackle the challenge of data heterogeneity in federated learning (FL), personalized FL has been proposed to maximize individual utility (model performance) by customizing personalized models for clients. Considering the significance of individual rationality, existing works have formulated clients’ participation decisions problem as hedonic games. However, they assume that clients can participate in only one collaborative coalition, constraining players’ attempts to join multiple coalitions. Different from prior works, we approach personalized FL from the perspective of hedonic overlapping coalition formation (OCF) games where rational clients can join multiple coalitions and generate their personalized model by weighting the local and coalition models. Nevertheless, the key challenge in analyzing the game is how to achieve a stable coalition structure where no clients would deviate from the current structure. This leads to our main question: what does a stable OCF structure look like? To address this problem, we first investigate the linear FL models for theoretical insights. Then, we design a heuristic algorithm for achieving an individually stable OCF structure. Experimental results demonstrate the feasibility of our algorithm for both linear and non-linear models, and show that our mechanism can improve the personalized model performance by up to 19% over existing methods.

Duke Scholars

Author Bing Luo DKU Faculty