Towards More Efficient Data Valuation in Healthcare Federated Learning using Ensembling
This addresses the need for fast and accurate data valuation in medical imaging federated learning, where institutional heterogeneity is common, but it is incremental as it builds on existing Shapley value approximation methods.
The paper tackles the problem of efficiently computing Shapley values for data valuation in healthcare federated learning, proposing SaFE, which achieves values close to exact Shapley values and outperforms existing approximations.
Federated Learning (FL) wherein multiple institutions collaboratively train a machine learning model without sharing data is becoming popular. Participating institutions might not contribute equally, some contribute more data, some better quality data or some more diverse data. To fairly rank the contribution of different institutions, Shapley value (SV) has emerged as the method of choice. Exact SV computation is impossibly expensive, especially when there are hundreds of contributors. Existing SV computation techniques use approximations. However, in healthcare where the number of contributing institutions are likely not of a colossal scale, computing exact SVs is still exorbitantly expensive, but not impossible. For such settings, we propose an efficient SV computation technique called SaFE (Shapley Value for Federated Learning using Ensembling). We empirically show that SaFE computes values that are close to exact SVs, and that it performs better than current SV approximations. This is particularly relevant in medical imaging setting where widespread heterogeneity across institutions is rampant and fast accurate data valuation is required to determine the contribution of each participant in multi-institutional collaborative learning.