Defending Against Patch-based Backdoor Attacks on Self-Supervised Learning

被引:5
|
作者
Tejankar, Ajinkya [1 ]
Sanjabi, Maziar [2 ]
Wang, Qifan [2 ]
Wang, Sinong [2 ]
Firooz, Hamed [2 ]
Pirsiavash, Hamed [1 ]
Tan, Liang [2 ]
机构
[1] Univ Calif Davis, Davis, CA 95616 USA
[2] Meta AI, Delaware, OH USA
关键词
D O I
10.1109/CVPR52729.2023.01178
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Recently, self-supervised learning (SSL) was shown to be vulnerable to patch-based data poisoning backdoor attacks. It was shown that an adversary can poison a small part of the unlabeled data so that when a victim trains an SSL model on it, the final model will have a backdoor that the adversary can exploit. This work aims to defend self-supervised learning against such attacks. We use a three-step defense pipeline, where we first train a model on the poisoned data. In the second step, our proposed defense algorithm (PatchSearch) uses the trained model to search the training data for poisoned samples and removes them from the training set. In the third step, a final model is trained on the cleaned-up training set. Our results show that PatchSearch is an effective defense. As an example, it improves a model's accuracy on images containing the trigger from 38.2% to 63.7% which is very close to the clean model's accuracy, 64.6%. Moreover, we show that PatchSearch outperforms baselines and state-of-the-art defense approaches including those using additional clean, trusted data. Our code is available at https://github.com/UCDvision/PatchSearch
引用
收藏
页码:12239 / 12249
页数:11
相关论文
共 50 条
  • [1] Backdoor Attacks on Self-Supervised Learning
    Saha, Aniruddha
    Tejankar, Ajinkya
    Koohpayegani, Soroush Abbasi
    Pirsiavash, Hamed
    2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2022, : 13327 - 13336
  • [2] BADFSS: Backdoor Attacks on Federated Self-Supervised Learning
    Zhang, Jiale
    Zhu, Chengcheng
    Di Wu
    Sun, Xiaobing
    Yong, Jianming
    Long, Guodong
    PROCEEDINGS OF THE THIRTY-THIRD INTERNATIONAL JOINT CONFERENCE ON ARTIFICIAL INTELLIGENCE, IJCAI 2024, 2024, : 548 - 558
  • [3] Invisible Backdoor Attacks on Key Regions Based on Target Neurons in Self-Supervised Learning
    Qian, Xiangyun
    He, Yusheng
    Zhang, Rui
    Kang, Zi
    Sheng, Yilin
    Xia, Hui
    KNOWLEDGE SCIENCE, ENGINEERING AND MANAGEMENT, PT III, KSEM 2024, 2024, 14886 : 122 - 134
  • [4] BadEncoder: Backdoor Attacks to Pre-trained Encoders in Self-Supervised Learning
    Jia, Jinyuan
    Liu, Yupei
    Gong, Neil Zhenqiang
    43RD IEEE SYMPOSIUM ON SECURITY AND PRIVACY (SP 2022), 2022, : 2043 - 2059
  • [5] On the Difficulty of Defending Self-Supervised Learning against Model Extraction
    Dziedzic, Adam
    Dhawan, Nikita
    Kaleem, Muhammad Ahmad
    Guan, Jonas
    Papernot, Nicolas
    INTERNATIONAL CONFERENCE ON MACHINE LEARNING, VOL 162, 2022,
  • [6] FedPD: Defending federated prototype learning against backdoor attacks
    Tan, Zhou
    Cai, Jianping
    Li, De
    Lian, Puwei
    Liu, Ximeng
    Che, Yan
    NEURAL NETWORKS, 2025, 184
  • [7] RoPE: Defending against backdoor attacks in federated learning systems
    Wang, Yongkang
    Zhai, Di-Hua
    Xia, Yuanqing
    KNOWLEDGE-BASED SYSTEMS, 2024, 293
  • [8] DEFENDING AGAINST BACKDOOR ATTACKS IN FEDERATED LEARNING WITH DIFFERENTIAL PRIVACY
    Miao, Lu
    Yang, Wei
    Hu, Rong
    Li, Lu
    Huang, Liusheng
    2022 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP), 2022, : 2999 - 3003
  • [9] Defending Against Data and Model Backdoor Attacks in Federated Learning
    Wang, Hao
    Mu, Xuejiao
    Wang, Dong
    Xu, Qiang
    Li, Kaiju
    IEEE INTERNET OF THINGS JOURNAL, 2024, 11 (24): : 39276 - 39294
  • [10] An Embarrassingly Simple Backdoor Attack on Self-supervised Learning
    Li, Changjiang
    Pang, Ren
    Xi, Zhaohan
    Du, Tianyu
    Ji, Shouling
    Yao, Yuan
    Wang, Ting
    2023 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION, ICCV, 2023, : 4344 - 4355