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 2305.11997stat.MLcs.AIcs.CYcs.ITcs.LGmath.IT
收藏

具有概率保证的神经网络鲁棒反事实解释

Robust Counterfactual Explanations for Neural Networks With Probabilistic Guarantees

University of Maryland(马里兰大学);JP Morgan Chase AI Research(摩根大通人工智能研究院)

Faisal Hamman, Erfaun Noorani, Saumitra Mishra, Daniele Magazzeni, Sanghamitra Dutta

AI总结:

针对模型更新后反事实解释失效的问题,提出“自然发生”模型变化抽象和“稳定性”度量,证明高稳定性反事实以高概率保持有效,并给出实用松弛方法。

摘要:

近年来,生成鲁棒的反事实解释引起了越来越多的关注,这类解释在模型被更新或发生轻微变化时仍能保持有效。在寻找鲁棒反事实的过程中,现有文献通常假设原始模型 $m$ 和新模型 $M$ 在参数空间上有界,即 $\|\ ext{Params}(M){-}\ ext{Params}(m)\|{<}Δ$。然而,模型在参数空间上可能发生显著变化,但其在给定数据集上的预测或准确率却几乎不变。在本工作中,我们引入了一种数学抽象,称为“自然发生”的模型变化,该抽象允许参数空间中的任意变化,但要求这些变化对位于数据流形上的点的预测影响有限。接下来,我们提出了一种度量——称为“稳定性”——用于量化可微模型(如神经网络)的反事实对潜在模型变化的鲁棒性。我们的主要贡献在于证明:根据我们的度量定义,具有足够高“稳定性”值的反事实在潜在的“自然发生”模型变化后,将以高概率保持有效(利用独立高斯变量的Lipschitz函数的集中界)。由于我们的量化依赖于数据点周围的局部Lipschitz常数,而该常数并非总是可得,我们还考察了所提出度量的实际松弛形式,并通过实验展示了如何将这些松弛形式用于寻找神经网络中接近、现实且在潜在模型变化后仍保持有效的鲁棒反事实。这项工作还与模型多重性(即Rashomon效应)有着有趣的关联。

英文摘要:

There is an emerging interest in generating robust counterfactual explanations that would remain valid if the model is updated or changed even slightly. Towards finding robust counterfactuals, existing literature often assumes that the original model $m$ and the new model $M$ are bounded in the parameter space, i.e., $\|\text{Params}(M){-}\text{Params}(m)\|{<}Δ$. However, models can often change significantly in the parameter space with little to no change in their predictions or accuracy on the given dataset. In this work, we introduce a mathematical abstraction termed $\textit{naturally-occurring}$ model change, which allows for arbitrary changes in the parameter space such that the change in predictions on points that lie on the data manifold is limited. Next, we propose a measure -- that we call $\textit{Stability}$ -- to quantify the robustness of counterfactuals to potential model changes for differentiable models, e.g., neural networks. Our main contribution is to show that counterfactuals with sufficiently high value of $\textit{Stability}$ as defined by our measure will remain valid after potential $\textit{naturally-occurring}$ model changes with high probability (leveraging concentration bounds for Lipschitz function of independent Gaussians). Since our quantification depends on the local Lipschitz constant around a data point which is not always available, we also examine practical relaxations of our proposed measure and demonstrate experimentally how they can be incorporated to find robust counterfactuals for neural networks that are close, realistic, and remain valid after potential model changes. This work also has interesting connections with model multiplicity, also known as, the Rashomon effect.

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。

↑