arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2409.05208cs.LGcs.AI
收藏

基于影响的归因可被操纵

Influence-based Attributions can be Manipulated

University of California, San Diego(加利福尼亚大学圣迭戈分校)

Chhavi Yadav, Ruihan Wu, Kamalika Chaudhuri

AI总结:

该研究发现基于影响函数的归因可被敌手系统性操纵,针对ResNet特征嵌入和标准表格公平性数据集上的逻辑回归模型提出了高效攻击方法,对对抗环境下这类归因的可靠性提出了质疑。

摘要:

影响函数(Influence Functions)是一种以有原则的方式将预测归因于训练数据的标准工具,广泛应用于数据估值、公平性等场景。本研究提出了操纵基于影响的归因的现实动机,并探究这类归因是否能被敌手系统性篡改。我们证实,对于在ResNet特征嵌入和标准表格公平性数据集上训练的逻辑回归模型,此类篡改确实可行,还提供了具备反向友好实现的高效攻击方法。本研究对对抗环境下基于影响的归因的可靠性提出了质疑。代码可在https://github.com/infinite-pursuits/influence-based-attributions-can-be-manipulated获取。

英文摘要:

Influence Functions are a standard tool for attributing predictions to training data in a principled manner and are widely used in applications such as data valuation and fairness. In this work, we present realistic incentives to manipulate influence-based attributions and investigate whether these attributions can be \textit{systematically} tampered by an adversary. We show that this is indeed possible for logistic regression models trained on ResNet feature embeddings and standard tabular fairness datasets and provide efficient attacks with backward-friendly implementations. Our work raises questions on the reliability of influence-based attributions in adversarial circumstances. Code is available at : \url{https://github.com/infinite-pursuits/influence-based-attributions-can-be-manipulated}

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。

↑