FedCCEA : A Practical Approach of Client Contribution Evaluation for Federated Learning
This addresses client selection and incentive allocation in federated learning, offering a practical solution for scenarios with limited client information, though it appears incremental as it builds on existing data valuation methods.
The paper tackles the problem of evaluating client contributions in federated learning under data accessibility restrictions by proposing FedCCEA, which uses an Accuracy Approximation Model to estimate contributions based on data size and quality, achieving precise estimation in non-IID settings and feasible evaluation time.
Client contribution evaluation, also known as data valuation, is a crucial approach in federated learning(FL) for client selection and incentive allocation. However, due to restrictions of accessibility of raw data, only limited information such as local weights and local data size of each client is open for quantifying the client contribution. Using data size from available information, we introduce an empirical evaluation method called Federated Client Contribution Evaluation through Accuracy Approximation(FedCCEA). This method builds the Accuracy Approximation Model(AAM), which estimates a simulated test accuracy using inputs of sampled data size and extracts the clients' data quality and data size to measure client contribution. FedCCEA strengthens some advantages: (1) enablement of data size selection to the clients, (2) feasible evaluation time regardless of the number of clients, and (3) precise estimation in non-IID settings. We demonstrate the superiority of FedCCEA compared to previous methods through several experiments: client contribution distribution, client removal, and robustness test to partial participation.