State Backdoor: Towards Stealthy Real-world Poisoning Attack on Vision-Language-Action Model in State Space
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911359997837312 |
|---|---|
| author | Guo, Ji Jiang, Wenbo Lin, Yansong Liu, Yijing Zhang, Ruichen Lu, Guomin Chen, Aiguo Han, Xinshuo Li, Hongwei Niyato, Dusit |
| author_facet | Guo, Ji Jiang, Wenbo Lin, Yansong Liu, Yijing Zhang, Ruichen Lu, Guomin Chen, Aiguo Han, Xinshuo Li, Hongwei Niyato, Dusit |
| contents | Vision-Language-Action (VLA) models are widely deployed in safety-critical embodied AI applications such as robotics. However, their complex multimodal interactions also expose new security vulnerabilities. In this paper, we investigate a backdoor threat in VLA models, where malicious inputs cause targeted misbehavior while preserving performance on clean data. Existing backdoor methods predominantly rely on inserting visible triggers into visual modality, which suffer from poor robustness and low insusceptibility in real-world settings due to environmental variability. To overcome these limitations, we introduce the State Backdoor, a novel and practical backdoor attack that leverages the robot arm's initial state as the trigger. To optimize trigger for insusceptibility and effectiveness, we design a Preference-guided Genetic Algorithm (PGA) that efficiently searches the state space for minimal yet potent triggers. Extensive experiments on five representative VLA models and five real-world tasks show that our method achieves over 90% attack success rate without affecting benign task performance, revealing an underexplored vulnerability in embodied AI systems. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_04266 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | State Backdoor: Towards Stealthy Real-world Poisoning Attack on Vision-Language-Action Model in State Space Guo, Ji Jiang, Wenbo Lin, Yansong Liu, Yijing Zhang, Ruichen Lu, Guomin Chen, Aiguo Han, Xinshuo Li, Hongwei Niyato, Dusit Cryptography and Security Machine Learning Vision-Language-Action (VLA) models are widely deployed in safety-critical embodied AI applications such as robotics. However, their complex multimodal interactions also expose new security vulnerabilities. In this paper, we investigate a backdoor threat in VLA models, where malicious inputs cause targeted misbehavior while preserving performance on clean data. Existing backdoor methods predominantly rely on inserting visible triggers into visual modality, which suffer from poor robustness and low insusceptibility in real-world settings due to environmental variability. To overcome these limitations, we introduce the State Backdoor, a novel and practical backdoor attack that leverages the robot arm's initial state as the trigger. To optimize trigger for insusceptibility and effectiveness, we design a Preference-guided Genetic Algorithm (PGA) that efficiently searches the state space for minimal yet potent triggers. Extensive experiments on five representative VLA models and five real-world tasks show that our method achieves over 90% attack success rate without affecting benign task performance, revealing an underexplored vulnerability in embodied AI systems. |
| title | State Backdoor: Towards Stealthy Real-world Poisoning Attack on Vision-Language-Action Model in State Space |
| topic | Cryptography and Security Machine Learning |
| url | https://arxiv.org/abs/2601.04266 |