arXiv cs.AI· R. Teal Witter, Yurong Liu, Christopher Musco·· 6 天前AI 评分23
回归调整的蒙特卡洛估计器：更精确估算 Shapley 值与概率值
Regression-adjusted Monte Carlo Estimators for Shapley Values and Probabilistic Values
AI 导读
研究者提出一种将蒙特卡洛采样与线性回归结合的新方法，可用 XGBoost 等任意函数族替代线性回归，并通过交叉拟合产生无偏估计。在八个数据集上，该方法估算 Shapley 值的误差比 Permutation SHAP 低 11 倍、比 Kernel SHAP 低 6.5 倍、比 Leverage SHAP 低 3.4 倍；对更一般的概率值，误差比此前最佳估计器低 390 倍。
正文 · 原文
Abstract:With origins in game theory, probabilistic values like Shapley values, Banzhaf values, and semi-values have emerged as a central tool in explainable AI. They are used for feature attribution, data attribution, data valuation, and more. Since exact computation can require exponential time, research has focused on efficient approximation methods using two techniques: Monte Carlo sampling and linear regression formulations. In this work, we present a new way of combining both of these techniques. Our approach is more flexible than prior algorithms, allowing for linear regression to be replaced with any function family whose probabilistic values can be computed efficiently. This allows us to harness the accuracy of tree-based models like XGBoost; with cross-fitting, the method also produces unbiased estimates. From experiments across eight datasets, we find that our methods give state-of-the-art performance for estimating probabilistic values. For Shapley values, the error of our methods can be $11\times$ lower than Permutation SHAP (the most popular Monte Carlo method), $6.5\times$ lower than Kernel SHAP (the most popular linear regression method), and $3.4\times$ lower than Leverage SHAP (the prior state-of-the-art Shapley value estimator). For more general probabilistic values, we can obtain error $390\times$ lower than the best estimator from prior work.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2506.11849 [cs.LG] |
| (or arXiv:2506.11849v3 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2506.11849 arXiv-issued DOI via DataCite |
Submission history
From: R. Teal Witter [view email]
[v1]
Fri, 13 Jun 2025 14:57:38 UTC (1,276 KB)
[v2]
Tue, 13 Jan 2026 02:37:05 UTC (1,280 KB)
[v3]
Thu, 8 Oct 2026 23:11:50 UTC (698 KB)
来源：arXiv cs.AI · arxiv.org