Influence-based Attributions can be Manipulated
This work highlights a vulnerability in widely used influence functions for data valuation and fairness, raising reliability concerns in adversarial settings.
The paper tackles the problem of adversarial manipulation of influence-based attributions, showing that these attributions can be systematically tampered with in logistic regression models on ResNet features and fairness datasets, with efficient attacks demonstrated.
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}