Saved in:
Bibliographic Details
Main Authors: Chen, Zhizhen, Zhao, Zhengyu, Dutta, Subrat Kishore, Lin, Chenhao, Shen, Chao, Zhang, Xiao
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2412.03908
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913960056324096
author Chen, Zhizhen
Zhao, Zhengyu
Dutta, Subrat Kishore
Lin, Chenhao
Shen, Chao
Zhang, Xiao
author_facet Chen, Zhizhen
Zhao, Zhengyu
Dutta, Subrat Kishore
Lin, Chenhao
Shen, Chao
Zhang, Xiao
contents Targeted data poisoning (TDP) aims to compromise the model's prediction on a specific (test) target by perturbing a small subset of training data. Existing work on TDP has focused on an overly ideal threat model in which the same image sample of the target is used during both poisoning and inference stages. However, in the real world, a target object often appears in complex variations due to changes of physical settings such as viewpoint, background, and lighting conditions. In this work, we take the first step toward understanding the real-world threats of TDP by studying its generalizability across varying physical conditions. In particular, we observe that solely optimizing gradient directions, as adopted by the best previous TDP method, achieves limited generalization. To address this limitation, we propose optimizing both the gradient direction and magnitude for more generalizable gradient matching, thereby leading to higher poisoning success rates. For instance, our method outperforms the state of the art by 19.49% when poisoning CIFAR-10 images targeting multi-view cars.
format Preprint
id arxiv_https___arxiv_org_abs_2412_03908
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Generalizable Targeted Data Poisoning against Varying Physical Objects
Chen, Zhizhen
Zhao, Zhengyu
Dutta, Subrat Kishore
Lin, Chenhao
Shen, Chao
Zhang, Xiao
Computer Vision and Pattern Recognition
Cryptography and Security
Machine Learning
Targeted data poisoning (TDP) aims to compromise the model's prediction on a specific (test) target by perturbing a small subset of training data. Existing work on TDP has focused on an overly ideal threat model in which the same image sample of the target is used during both poisoning and inference stages. However, in the real world, a target object often appears in complex variations due to changes of physical settings such as viewpoint, background, and lighting conditions. In this work, we take the first step toward understanding the real-world threats of TDP by studying its generalizability across varying physical conditions. In particular, we observe that solely optimizing gradient directions, as adopted by the best previous TDP method, achieves limited generalization. To address this limitation, we propose optimizing both the gradient direction and magnitude for more generalizable gradient matching, thereby leading to higher poisoning success rates. For instance, our method outperforms the state of the art by 19.49% when poisoning CIFAR-10 images targeting multi-view cars.
title Generalizable Targeted Data Poisoning against Varying Physical Objects
topic Computer Vision and Pattern Recognition
Cryptography and Security
Machine Learning
url https://arxiv.org/abs/2412.03908