{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/vulnerability-of-deep-reinforcement-learning","title":"Vulnerability of Deep Reinforcement Learning to Policy Induction Attacks","arxiv_id":"1701.04143","date":"2017-01-16","proceeding":null,"authors":["Vahid Behzadan","Arslan Munir"],"abstract":"Deep learning classifiers are known to be inherently vulnerable to\nmanipulation by intentionally perturbed inputs, named adversarial examples. In\nthis work, we establish that reinforcement learning techniques based on Deep\nQ-Networks (DQNs) are also vulnerable to adversarial input perturbations, and\nverify the transferability of adversarial examples across different DQN models.\nFurthermore, we present a novel class of attacks based on this vulnerability\nthat enable policy manipulation and induction in the learning process of DQNs.\nWe propose an attack mechanism that exploits the transferability of adversarial\nexamples to implement policy induction attacks on DQNs, and demonstrate its\nefficacy and impact through experimental study of a game-learning scenario.","url_abs":"http://arxiv.org/abs/1701.04143v1","url_pdf":"http://arxiv.org/pdf/1701.04143v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"vulnerability-of-deep-reinforcement-learning","repo_url":"https://github.com/coderatwork7/attack","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}}],"tasks":[{"task_slug":"deep-reinforcement-learning","task_name":"Deep Reinforcement Learning"},{"task_slug":"reinforcement-learning","task_name":"Reinforcement Learning"},{"task_slug":"reinforcement-learning-1","task_name":"Reinforcement Learning (RL)"},{"task_slug":"reinforcement-learning-2","task_name":"reinforcement-learning"}],"methods":[{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"dqn","method_name":"DQN"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"q-learning","method_name":"Q-Learning"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1701.04143","atlas_url":"https://app.syntology.ai/?focus=1701.04143","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}