{"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/robust-adversarial-reinforcement-learning","title":"Robust Adversarial Reinforcement Learning","arxiv_id":"1703.02702","date":"2017-03-08","proceeding":"ICML 2017 8","authors":["Lerrel Pinto","James Davidson","Rahul Sukthankar","Abhinav Gupta"],"abstract":"Deep neural networks coupled with fast simulation and improved computation\nhave led to recent successes in the field of reinforcement learning (RL).\nHowever, most current RL-based approaches fail to generalize since: (a) the gap\nbetween simulation and real world is so large that policy-learning approaches\nfail to transfer; (b) even if policy learning is done in real world, the data\nscarcity leads to failed generalization from training to test scenarios (e.g.,\ndue to different friction or object masses). Inspired from H-infinity control\nmethods, we note that both modeling errors and differences in training and test\nscenarios can be viewed as extra forces/disturbances in the system. This paper\nproposes the idea of robust adversarial reinforcement learning (RARL), where we\ntrain an agent to operate in the presence of a destabilizing adversary that\napplies disturbance forces to the system. The jointly trained adversary is\nreinforced -- that is, it learns an optimal destabilization policy. We\nformulate the policy learning as a zero-sum, minimax objective function.\nExtensive experiments in multiple environments (InvertedPendulum, HalfCheetah,\nSwimmer, Hopper and Walker2d) conclusively demonstrate that our method (a)\nimproves training stability; (b) is robust to differences in training/test\nconditions; and c) outperform the baseline even in the absence of the\nadversary.","url_abs":"http://arxiv.org/abs/1703.02702v1","url_pdf":"http://arxiv.org/pdf/1703.02702v1.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":"robust-adversarial-reinforcement-learning","repo_url":"https://github.com/Jekyll1021/RARL","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"robust-adversarial-reinforcement-learning","repo_url":"https://github.com/Yangangren/Adversarial-Reinforcement-learning","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}},{"paper_slug":"robust-adversarial-reinforcement-learning","repo_url":"https://github.com/Yangangren/Adversarial_RL","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}},{"paper_slug":"robust-adversarial-reinforcement-learning","repo_url":"https://github.com/davidsonic/robust-grasp","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"robust-adversarial-reinforcement-learning","repo_url":"https://github.com/lerrel/gym-adv","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"robust-adversarial-reinforcement-learning","repo_url":"https://github.com/lerrel/rllab-adv","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"robust-adversarial-reinforcement-learning","repo_url":"https://github.com/sureli/rrls","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"friction","task_name":"Friction"},{"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":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1703.02702","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"1703.02702"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/davidsonic/robust-grasp","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/Jekyll1021/RARL","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/Yangangren/Adversarial-Reinforcement-learning","reach":{"status":"ok"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/lerrel/rllab-adv","reach":{"status":"ok"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/lerrel/gym-adv","reach":{"status":"ok"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/Yangangren/Adversarial_RL","reach":{"status":"ok"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/sureli/rrls","reach":{"status":"ok","spdx":"MIT"}}],"summary":{"ran_honours":2,"unverified":2},"by_repo_kind":{"listed":{"samples":4,"ran":2,"repositories":2}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":4,"samples":[{"code_sha256_prefix":"0173e4280d9a9884","entry":"flatten_lists","repo":"Jekyll1021/RARL","repo_kind":"listed","path":"reimplementation/PPO_RARL.py","file_url":"https://github.com/Jekyll1021/RARL/blob/HEAD/reimplementation/PPO_RARL.py","link_basis":"first_harvest_node","language":"python","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"0173e4280d9a9884"}},{"code_sha256_prefix":"d30a9422beea45f4","entry":"test","repo":"davidsonic/robust-grasp","repo_kind":"listed","path":"src/PPO_RARL.py","file_url":"https://github.com/davidsonic/robust-grasp/blob/HEAD/src/PPO_RARL.py","link_basis":"first_harvest_node","language":"python","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"d30a9422beea45f4"}},{"code_sha256_prefix":"a2029b506300c864","entry":"inject_noise","repo":"Jekyll1021/RARL","repo_kind":"listed","path":"extend/RARL_based_interaction.py","file_url":"https://github.com/Jekyll1021/RARL/blob/HEAD/extend/RARL_based_interaction.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"a2029b506300c864"}},{"code_sha256_prefix":"7c443f61be7575f7","entry":"train_baseline_model","repo":"Jekyll1021/RARL","repo_kind":"listed","path":"extend/RARL_based_interaction.py","file_url":"https://github.com/Jekyll1021/RARL/blob/HEAD/extend/RARL_based_interaction.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"7c443f61be7575f7"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}