基于正则化的竞争:Shapley值与最小充分子集
The Struggles of Feature-Based Explanations: Shapley Values vs. Minimal Sufficient Subsets
摘要:
为了让神经模型获得广泛的公众信任并确保公平性,我们必须对它们的预测做出人类可以理解的解释。最近,越来越多的作品集中在解释神经模型的预测输入功能的相关性。在这项工作中,我们表明,基于特征的解释带来的问题,即使是解释琐碎的模型。我们表明,在某些情况下,至少存在两个基于地面实况特征的解释,而且,有时,它们都不足以提供模型决策过程的完整视图。此外,我们表明,两个流行的类的解释者,沙普利解释者和最小的充分子集解释者,目标从根本上不同类型的地面真理的解释,尽管显然隐含的假设,解释者应该寻找一个特定的功能为基础的解释。这些发现为解释者的发展和选择提供了一个新的维度。
英文摘要:
For neural models to garner widespread public trust and ensure fairness, we must have human-intelligible explanations for their predictions. Recently, an increasing number of works focus on explaining the predictions of neural models in terms of the relevance of the input features. In this work, we show that feature-based explanations pose problems even for explaining trivial models. We show that, in certain cases, there exist at least two ground-truth feature-based explanations, and that, sometimes, neither of them is enough to provide a complete view of the decision-making process of the model. Moreover, we show that two popular classes of explainers, Shapley explainers and minimal sufficient subsets explainers, target fundamentally different types of ground-truth explanations, despite the apparently implicit assumption that explainers should look for one specific feature-based explanation. These findings bring an additional dimension to consider in both developing and choosing explainers.
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。