How to Evaluate Participant Contributions in Decentralized Federated Learning
This addresses the need for fair incentive mechanisms in DFL systems, which is crucial for encouraging active participation and transparency, though it is an incremental improvement over existing centralized FL methods.
The paper tackles the problem of evaluating participant contributions in decentralized federated learning (DFL) by proposing TRIP-Shapley, a method that traces round-wise local contributions to estimate overall contributions without collecting models, achieving results close to ground-truth Shapley values and demonstrating scalability and robustness to dishonest clients.
Federated learning (FL) enables multiple clients to collaboratively train machine learning models without sharing local data. In particular, decentralized FL (DFL), where clients exchange models without a central server, has gained attention for mitigating communication bottlenecks. Evaluating participant contributions is crucial in DFL to incentivize active participation and enhance transparency. However, existing contribution evaluation methods for FL assume centralized settings and cannot be applied directly to DFL due to two challenges: the inaccessibility of each client to non-neighboring clients' models, and the necessity to trace how contributions propagate in conjunction with peer-to-peer model exchanges over time. To address these challenges, we propose TRIP-Shapley, a novel contribution evaluation method for DFL. TRIP-Shapley formulates the clients' overall contributions by tracing the propagation of the round-wise local contributions. In this way, TRIP-Shapley accurately reflects the delayed and gradual influence propagation, as well as allowing a lightweight coordinator node to estimate the overall contributions without collecting models, but based solely on locally observable contributions reported by each client. Experiments demonstrate that TRIP-Shapley is sufficiently close to the ground-truth Shapley value, is scalable to large-scale scenarios, and remains robust in the presence of dishonest clients.