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 2401.06465cs.AIcs.LGstat.ME
收藏

健全性检查再探:对模型参数随机化检验修复的探索

Sanity Checks Revisited: An Exploration to Repair the Model Parameter Randomisation Test

BIFOLD – Berlin Institute for the Foundations of Learning and Data(柏林学习与数据基础研究所(BIFOLD));Department of Electrical Engineering and Computer Science, TU Berlin(柏林工业大学电气工程与计算机科学系);Department of Artificial Intelligence, Fraunhofer Heinrich Hertz Institute(弗劳恩霍夫海因里希·赫兹研究所人工智能系);UMI Lab, Leibniz Institute of Agricultural Engineering and Bioeconomy e.V. (ATB)(莱布尼茨农业工程与生物经济研究所UMI实验室);Department of Computer Science, University of Potsdam(波茨坦大学计算机科学系)

Anna Hedström, Leander Weber, Sebastian Lapuschkin, Marina MC Höhne

AI总结:

本文针对模型参数随机化检验(MPRT)的方法论缺陷,提出平滑MPRT和高效MPRT两种改进,通过降噪采样和基于解释复杂度上升的重新解释提升度量可靠性,从而增强XAI方法应用的可信度。

摘要:

模型参数随机化检验(MPRT)在可解释人工智能(XAI)社区中被广泛认可,其评估原则具有良好动机:解释函数应对模型函数参数的变化保持敏感。然而,近期研究已识别出MPRT在经验解释上的若干方法论缺陷。为应对这些缺陷,我们对原始MPRT引入两种改进——平滑MPRT和高效MPRT,前者通过采样最小化噪声对评估结果的影响,后者通过在完全参数随机化后重新以解释复杂度的上升来解释该检验,从而规避了对有偏相似性度量的依赖。我们的实验结果表明,这些所提出的变体提升了度量可靠性,从而使得XAI方法的应用更加可信。

英文摘要:

The Model Parameter Randomisation Test (MPRT) is widely acknowledged in the eXplainable Artificial Intelligence (XAI) community for its well-motivated evaluative principle: that the explanation function should be sensitive to changes in the parameters of the model function. However, recent works have identified several methodological caveats for the empirical interpretation of MPRT. To address these caveats, we introduce two adaptations to the original MPRT -- Smooth MPRT and Efficient MPRT, where the former minimises the impact that noise has on the evaluation results through sampling and the latter circumvents the need for biased similarity measurements by re-interpreting the test through the explanation's rise in complexity, after full parameter randomisation. Our experimental results demonstrate that these proposed variants lead to improved metric reliability, thus enabling a more trustworthy application of XAI methods.

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。

↑