arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2502.12295cs.LGcs.CCcs.LO
收藏

论(多种)Shapley值的计算可处理性

On the Computational Tractability of the (Many) Shapley Values

Université de Nantes(南特大学);The Hebrew University of Jerusalem(耶路撒冷希伯来大学)

Reda Marzouk, Shahaf Bassan, Guy Katz, Colin de la Higuera

AI总结:

本文分析Conditional、Interventional和Baseline SHAP在局部与全局计算下的复杂度,证明HMM分布中多类ML模型可多项式求解,并给出神经网络和树集成上的不可处理性结论。

摘要:

近期研究考察了在各种模型和分布下计算Shapley加性解释(也称为SHAP)的计算复杂度,揭示了其在不同设置下的可处理性或不可处理性。然而,这些研究主要关注一种称为Conditional SHAP的特定变体,尽管还存在许多其他变体并解决不同的局限性。在本工作中,我们分析了计算范围更广的此类变体的复杂度,包括Conditional、Interventional和Baseline SHAP,同时探索局部和全局计算。我们表明,在Hidden Markov Model分布下,对于多种ML模型,局部和全局Interventional与Baseline SHAP均可在多项式时间内计算,从而将TreeSHAP等流行算法扩展到经验分布之外。不利的一面是,我们证明了这些变体在广泛的神经网络和树集成上存在不可处理性结果。我们相信,我们的结果强调了计算Shapley值的复杂多样性,展示了其复杂度如何由具体SHAP变体、模型类型和分布共同显著塑造。

英文摘要:

Recent studies have examined the computational complexity of computing Shapley additive explanations (also known as SHAP) across various models and distributions, revealing their tractability or intractability in different settings. However, these studies primarily focused on a specific variant called Conditional SHAP, though many other variants exist and address different limitations. In this work, we analyze the complexity of computing a much broader range of such variants, including Conditional, Interventional, and Baseline SHAP, while exploring both local and global computations. We show that both local and global Interventional and Baseline SHAP can be computed in polynomial time for various ML models under Hidden Markov Model distributions, extending popular algorithms such as TreeSHAP beyond empirical distributions. On the downside, we prove intractability results for these variants over a wide range of neural networks and tree ensembles. We believe that our results emphasize the intricate diversity of computing Shapley values, demonstrating how their complexity is substantially shaped by both the specific SHAP variant, the model type, and the distribution.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑