arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2305.13271stat.MLcs.LG
收藏

MAGDiff:通过深度神经网络的激活图进行协变量数据集偏移检测

MAGDiff: Covariate Data Set Shift Detection via Activation Graphs of Deep Neural Networks

Université Paris-Saclay(巴黎萨克雷大学);CNRS(法国国家科学研究中心);Inria(法国国家信息与自动化研究所);LIGM(信息与建模实验室);Université Gustave Eiffel(Gustave Eiffel大学);Fujitsu Limited(富士通公司);Institute of Mathematics for Industry, Kyushu University(九州大学产业数学研究所)

Charles Arnal, Felix Hensel, Mathieu Carrière, Théo Lacombe, Hiroaki Kurihara, Yuichi Ike, Frédéric Chazal

AI总结:

提出 MAGDiff 表示,利用神经网络在训练与目标分布样本上的激活图差异进行协变量偏移检测,无需额外训练模型,并在多数据集 KS 检验中显著优于基线。

摘要:

尽管神经网络已成功应用于多种任务,但与其他机器学习方法一样,它们仍受到对数据变化敏感性的限制:训练数据与部署数据之间的分布差异可能严重影响其性能。在本文中,我们提出了一系列名为 MAGDiff 的新型表示,可从任意给定的神经网络分类器中提取,无需训练专门用于该任务的新模型,即可高效检测协变量数据集偏移。这些表示通过比较神经网络在属于训练分布和目标分布的样本上的激活图来计算,并为常用于数据集偏移检测的双样本检验生成强大的、适配数据和任务的统计量。我们通过在多个不同数据集和偏移类型上测量双样本 Kolmogorov-Smirnov(KS)检验的统计功效进行了实证验证,结果表明,与依赖网络输出的先进基线相比,我们的新型表示带来了显著提升。

英文摘要:

Despite their successful application to a variety of tasks, neural networks remain limited, like other machine learning methods, by their sensitivity to shifts in the data: their performance can be severely impacted by differences in distribution between the data on which they were trained and that on which they are deployed. In this article, we propose a new family of representations, called MAGDiff, that we extract from any given neural network classifier and that allows for efficient covariate data shift detection without the need to train a new model dedicated to this task. These representations are computed by comparing the activation graphs of the neural network for samples belonging to the training distribution and to the target distribution, and yield powerful data- and task-adapted statistics for the two-sample tests commonly used for data set shift detection. We demonstrate this empirically by measuring the statistical powers of two-sample Kolmogorov-Smirnov (KS) tests on several different data sets and shift types, and showing that our novel representations induce significant improvements over a state-of-the-art baseline relying on the network output.

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。

↑