Hibernated Backdoor: A Mutual Information Empowered Backdoor Attack to Deep Neural Networks
Rui Ning, Jiang Li, Chunsheng Xin, Hongyi Wu, Chonggang Wang
Abstract
We report a new neural backdoor attack, named Hibernated Backdoor, which is stealthy, aggressive and devastating. The backdoor is planted in a hibernated mode to avoid being detected. Once deployed and fine-tuned on end-devices, the hibernated backdoor turns into the active state that can be exploited by the attacker. To the best of our knowledge, this is the first hibernated neural backdoor attack. It is achieved by maximizing the mutual information (MI) between the gradients of regular and malicious data on the model. We introduce a practical algorithm to achieve MI maximization to effectively plant the hibernated backdoor. To evade adaptive defenses, we further develop a targeted hibernated backdoor, which can only be activated by specific data samples and thus achieves a higher degree of stealthiness. We show the hibernated backdoor is robust and cannot be removed by existing backdoor removal schemes. It has been fully tested on four datasets with two neural network architectures, compared to five existing backdoor attacks, and evaluated using seven backdoor detection schemes. The experiments demonstrate the effectiveness of the hibernated backdoor attack under various settings.
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
7 top-tier papers have cited it since. Know what they found before your reviewers do.
Lune can draw on
the full paper · the top-tier papers that cite it · related work
Free to start. No credit card required.
Lineage
- Latent Backdoor Attacks on Deep Neural Networks cites Trojaning Attack on Neural Networks
- GangSweep: Sweep out Neural Backdoors by GAN cites Latent Backdoor Attacks on Deep Neural Networks
- GangSweep: Sweep out Neural Backdoors by GAN cites Trojaning Attack on Neural Networks
- GangSweep: Sweep out Neural Backdoors by GAN cites Hidden Trigger Backdoor Attacks
- BadRL: Sparse Targeted Backdoor Attack against Reinforcement Learning cites Hidden Trigger Backdoor Attacks
- UBA-Inf: Unlearning Activated Backdoor Attack with Influence-Driven Camouflage cites Trojaning Attack on Neural Networks
- Conditional Backdoor Attack via JPEG Compression cites Trojaning Attack on Neural Networks
- Taught Well Learned Ill: Towards Distillation-conditional Backdoor Attack cites Trojaning Attack on Neural Networks
Cited by top-tier papers7
- BadRL: Sparse Targeted Backdoor Attack against Reinforcement LearningJing Cui, Yufei Han, Yuzhe Ma, Jianbin Jiao et al.AAAI 2024 · 31 citations
- Conditional Backdoor Attack via JPEG CompressionQiuyu Duan, Zhongyun Hua, Qing Liao, Yushu Zhang et al.AAAI 2024 · 21 citations
- UBA-Inf: Unlearning Activated Backdoor Attack with Influence-Driven CamouflageZirui Huang, Yunlong Mao, Sheng ZhongUSENIX Security 2024 · 16 citations
- SADBA: Self-Adaptive Distributed Backdoor Attack Against Federated LearningJun Feng, Yuzhe Lai, Hong Sun, Bocheng RenAAAI 2025 · 9 citations
- Taught Well Learned Ill: Towards Distillation-conditional Backdoor AttackYukun Chen, Boheng Li, Yu Yuan, Leyi Qi et al.NeurIPS 2025 · 6 citations
Builds on7
- Trojaning Attack on Neural NetworksYingqi Liu, Shiqing Ma, Yousra Aafer, Wen-Chuan Lee et al.NDSS 2018 · 1,377 citations
- Hidden Trigger Backdoor AttacksAniruddha Saha, Akshayvarun Subramanya, Hamed PirsiavashAAAI 2020 · 743 citations
- Latent Backdoor Attacks on Deep Neural NetworksYuanshun Yao, Huiying Li, Haitao Zheng, Ben Y. ZhaoCCS 2019 · 465 citations
- Restricting the Flow: Information Bottlenecks for AttributionKarl Schulz, Leon Sixt, Federico Tombari, Tim LandgrafICLR 2020 · 220 citations
- GangSweep: Sweep out Neural Backdoors by GANLiuwan Zhu, Rui Ning, Cong Wang, Chunsheng Xin et al.ACM MM 2020 · 59 citations
Related papers
- Pre-activation Distributions Expose Backdoor NeuronsRunkai Zheng, Rongjun Tang, Jianze Li, Li LiuNeurIPS 2022 · 51 citations
- ReVeil: Unconstrained Concealed Backdoor Attack on Deep Neural Networks using Machine UnlearningManaar Alam, Hithem Lamri, Michail ManiatakosDAC 2025 · 3 citations
- Dormant Backdoor: Weaponizing Model Finetuning for Feasible Backdoor Attacks Against Pretrained ModelsRuitao Li, Jiakai Wang, Hairong Chen, Huihu Ding et al.AAAI 2026
- Beating Backdoor Attack at Its Own GameMin Liu, Alberto L. Sangiovanni-Vincentelli, Xiangyu YueICCV 2023 · 19 citations
- MM-BD: Post-Training Detection of Backdoor Attacks with Arbitrary Backdoor Pattern Types Using a Maximum Margin StatisticHang Wang, Zhen Xiang, David J. Miller, George KesidisS&P 2024 · 81 citations
