Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces Trajectory Shapley Value (TSV), a novel metric for evaluating client contributions in federated learning by assessing their impact on the global model's optimization trajectory using a validation-based utility function. Based on TSV, the authors propose FedTSV, an adaptive aggregation method that dynamically weights client updates based on their per-round TSV scores. Experiments on benchmark datasets demonstrate that FedTSV improves convergence speed, robustness against adversarial clients, and fairness in contribution assessment compared to traditional fixed-weight aggregation.
FedTSV dynamically re-weights client updates in federated learning based on their impact on the global model's optimization trajectory, leading to faster convergence, increased robustness, and fairer contribution assessments.
Federated learning is an emerging distributed paradigm that addresses the challenges posed by heterogeneous, privacy-sensitive data. It enables multiple clients to train a model collaboratively by aggregating their local updates at a server. However, conventional aggregation schemes typically use fixed weights that fail to reflect unequal and time-varying client contributions, leading to biased and unstable learning. To improve fairness and stability, we propose the Trajectory Shapley Value (TSV), a contribution metric that evaluates how each client influences the optimization trajectory of the global model using a validation-based, temporally consistent utility. Building on TSV, we design FedTSV, an adaptive aggregation method that converts per-round evaluations into dynamic client weights, allowing the server to respond to heterogeneous and adversarial participation in real time. Experiments on benchmark datasets show that FedTSV accelerates convergence, improves robustness, and yields more equitable contribution assessments, thereby providing a principled foundation for fairness-aware federated optimization.