Anirudh Rushil, Thiagarajan Jayaraman J, Sridhar Rahul, Bremer Peer-Timo
Center for Applied Scientific Computing (CASC), Lawrence Livermore National Laboratory, Livermore, CA, United States.
Walmart Labs, California, CA, United States.
Front Big Data. 2021 May 4;4:589417. doi: 10.3389/fdata.2021.589417. eCollection 2021.
Interpretability has emerged as a crucial aspect of building trust in machine learning systems, aimed at providing insights into the working of complex neural networks that are otherwise opaque to a user. There are a plethora of existing solutions addressing various aspects of interpretability ranging from identifying prototypical samples in a dataset to explaining image predictions or explaining mis-classifications. While all of these diverse techniques address seemingly different aspects of interpretability, we hypothesize that a large family of interepretability tasks are variants of the same central problem which is identifying change in a model's prediction. This paper introduces MARGIN, a simple yet general approach to address a large set of interpretability tasks MARGIN exploits ideas rooted in graph signal analysis to determine influential nodes in a graph, which are defined as those nodes that maximally describe a function defined on the graph. By carefully defining task-specific graphs and functions, we demonstrate that MARGIN outperforms existing approaches in a number of disparate interpretability challenges.
可解释性已成为机器学习系统中建立信任的关键方面，旨在深入了解复杂神经网络的工作方式，否则用户将对其一无所知。现有大量解决方案涉及可解释性的各个方面，从识别数据集中的原型样本到解释图像预测或错误分类。虽然所有这些不同的技术都解决了可解释性看似不同的方面，但我们假设一大类可解释性任务是同一个核心问题的变体，即识别模型预测中的变化。本文介绍了MARGIN，这是一种简单而通用的方法，用于解决大量可解释性任务。MARGIN利用源于图信号分析的思想来确定图中的有影响力的节点，这些节点被定义为那些最大程度描述图上定义的函数的节点。通过仔细定义特定任务的图和函数，我们证明了MARGIN在许多不同的可解释性挑战中优于现有方法。
