Towards Scalable Game-Theory-based Attribution
Seminar Series: Weijing Tang, CMU.
Shapley values and related probabilistic values provide a flexible model-agnostic framework for attributing the behavior of a black-box system to data points, features, or other components, with a wide range of applications from model explanation and data valuation to credit assignment and compensation. Their practical use, however, is limited by a fundamental computational challenge: exact evaluation requires considering exponentially many coalitions, while even a single coalition evaluation may involve expensive model training or inference. In this talk, Tang will introduce a unified first-order perspective on Monte Carlo estimation of probabilistic values. Although existing estimators are motivated by seemingly different strategies, many share a common error structure that reveals how coalition sampling and surrogate modeling jointly determine statistical efficiency. This perspective leads to an Efficiency-Aware Surrogate-adjusted Estimator (EASE) that directly minimizes an explicit first-order approximation to the mean squared error. We demonstrate that EASE consistently improves estimation accuracy compared to existing estimators under the same utility-evaluation budget. This is a joint work with Ziqi Liu, Kiljae Lee, and Yuan Zhang.
