Stealthy data poisoning attack method on offline reinforcement learning in unmanned systems
Aiming at the limitations in effectiveness and stealth of existing offline reinforcement learning(RL) data poisoning attacks, a critical time-step dynamic poisoning attack was proposed, perturbing important samples to achieve efficient and covert attacks. Temporal difference errors, identified throu...
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | zho |
| Published: |
Editorial Department of Journal on Communications
2024-12-01
|
| Series: | Tongxin xuebao |
| Subjects: | |
| Online Access: | http://www.joconline.com.cn/zh/article/doi/10.11959/j.issn.1000-436x.2024264/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832595467522801664 |
|---|---|
| author | ZHOU Xue MAN Dapeng XU Chen LYU Jiguang ZENG Fanyi GAO Chaoyang YANG Wu |
| author_facet | ZHOU Xue MAN Dapeng XU Chen LYU Jiguang ZENG Fanyi GAO Chaoyang YANG Wu |
| author_sort | ZHOU Xue |
| collection | DOAJ |
| description | Aiming at the limitations in effectiveness and stealth of existing offline reinforcement learning(RL) data poisoning attacks, a critical time-step dynamic poisoning attack was proposed, perturbing important samples to achieve efficient and covert attacks. Temporal difference errors, identified through theoretical analysis as crucial for model learning, were used to guide poisoning target selection. A bi-objective optimization approach was introduced to minimize perturbation magnitude while maximizing the negative impact on performance. Experimental results show that with only a 1% poisoning rate, the method reduces agent performance by 84%, revealing the sensitivity and vulnerability of offline RL models in unmanned systems. |
| format | Article |
| id | doaj-art-5654a38851454c37a0f434562d00aac5 |
| institution | Kabale University |
| issn | 1000-436X |
| language | zho |
| publishDate | 2024-12-01 |
| publisher | Editorial Department of Journal on Communications |
| record_format | Article |
| series | Tongxin xuebao |
| spelling | doaj-art-5654a38851454c37a0f434562d00aac52025-01-18T19:00:05ZzhoEditorial Department of Journal on CommunicationsTongxin xuebao1000-436X2024-12-0145162780268668Stealthy data poisoning attack method on offline reinforcement learning in unmanned systemsZHOU XueMAN DapengXU ChenLYU JiguangZENG FanyiGAO ChaoyangYANG WuAiming at the limitations in effectiveness and stealth of existing offline reinforcement learning(RL) data poisoning attacks, a critical time-step dynamic poisoning attack was proposed, perturbing important samples to achieve efficient and covert attacks. Temporal difference errors, identified through theoretical analysis as crucial for model learning, were used to guide poisoning target selection. A bi-objective optimization approach was introduced to minimize perturbation magnitude while maximizing the negative impact on performance. Experimental results show that with only a 1% poisoning rate, the method reduces agent performance by 84%, revealing the sensitivity and vulnerability of offline RL models in unmanned systems.http://www.joconline.com.cn/zh/article/doi/10.11959/j.issn.1000-436x.2024264/unmanned systemoffline reinforcement learningdata poisoning attackdata security |
| spellingShingle | ZHOU Xue MAN Dapeng XU Chen LYU Jiguang ZENG Fanyi GAO Chaoyang YANG Wu Stealthy data poisoning attack method on offline reinforcement learning in unmanned systems Tongxin xuebao unmanned system offline reinforcement learning data poisoning attack data security |
| title | Stealthy data poisoning attack method on offline reinforcement learning in unmanned systems |
| title_full | Stealthy data poisoning attack method on offline reinforcement learning in unmanned systems |
| title_fullStr | Stealthy data poisoning attack method on offline reinforcement learning in unmanned systems |
| title_full_unstemmed | Stealthy data poisoning attack method on offline reinforcement learning in unmanned systems |
| title_short | Stealthy data poisoning attack method on offline reinforcement learning in unmanned systems |
| title_sort | stealthy data poisoning attack method on offline reinforcement learning in unmanned systems |
| topic | unmanned system offline reinforcement learning data poisoning attack data security |
| url | http://www.joconline.com.cn/zh/article/doi/10.11959/j.issn.1000-436x.2024264/ |
| work_keys_str_mv | AT zhouxue stealthydatapoisoningattackmethodonofflinereinforcementlearninginunmannedsystems AT mandapeng stealthydatapoisoningattackmethodonofflinereinforcementlearninginunmannedsystems AT xuchen stealthydatapoisoningattackmethodonofflinereinforcementlearninginunmannedsystems AT lyujiguang stealthydatapoisoningattackmethodonofflinereinforcementlearninginunmannedsystems AT zengfanyi stealthydatapoisoningattackmethodonofflinereinforcementlearninginunmannedsystems AT gaochaoyang stealthydatapoisoningattackmethodonofflinereinforcementlearninginunmannedsystems AT yangwu stealthydatapoisoningattackmethodonofflinereinforcementlearninginunmannedsystems |