Faithful Group Shapley Value

Ziqi Liu (Carnegie Mellon University) · Yuan Zhang (Ohio State University, Columbus) · Weijing Tang (Carnegie Mellon University) · Kiljae Lee (The Ohio State University)

approximation accuracyapproximation algorithm computational efficiencydata providersdata shapley data valuation empirical experimentsfaithful group shapley valuefgsvgroup-level valuationmachine learning modelsmathematical insightsshell company attacksstate-of-the-art methodsstrategic group splitting

Data Shapley is an important tool for data valuation, which quantifies the contribution of individual data points to machine learning models. In practice, group-level data valuation is desirable when data providers contribute data in batch. However, we identify that existing group-level extensions of Data Shapley are vulnerable to \emph{shell company attacks}, where strategic group splitting can unfairly inflate valuations. We propose Faithful Group Shapley Value (FGSV) that uniquely defends against such attacks. Building on original mathematical insights, we develop a provably fast and accurate approximation algorithm for computing FGSV. Empirical experiments demonstrate that our algorithm significantly outperforms state-of-the-art methods in computational efficiency and approximation accuracy, while ensuring faithful group-level valuation.