{"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/reducing-exploitability-with-population-based","title":"Reducing Exploitability with Population Based Training","arxiv_id":"2208.05083","date":"2022-08-10","proceeding":null,"authors":["Pavel Czempin","Adam Gleave"],"abstract":"Self-play reinforcement learning has achieved state-of-the-art, and often superhuman, performance in a variety of zero-sum games. Yet prior work has found that policies that are highly capable against regular opponents can fail catastrophically against adversarial policies: an opponent trained explicitly against the victim. Prior defenses using adversarial training were able to make the victim robust to a specific adversary, but the victim remained vulnerable to new ones. We conjecture this limitation was due to insufficient diversity of adversaries seen during training. We analyze a defense using population based training to pit the victim against a diverse set of opponents. We evaluate this defense's robustness against new adversaries in two low-dimensional environments. This defense increases robustness against adversaries, as measured by the number of attacker training timesteps to exploit the victim. Furthermore, we show that robustness is correlated with the size of the opponent population.","url_abs":"https://arxiv.org/abs/2208.05083v3","url_pdf":"https://arxiv.org/pdf/2208.05083v3.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":"reducing-exploitability-with-population-based","repo_url":"https://github.com/humancompatibleai/reducing-exploitability","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"diversity","task_name":"Diversity"}],"methods":[{"method_slug":"population-based-training","method_name":"Population Based Training"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2208.05083","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2208.05083"}},"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/humancompatibleai/reducing-exploitability","reach":{"status":"ok","spdx":"MIT"}}],"summary":{"unverified":5},"by_repo_kind":{"official":{"samples":5,"ran":0,"repositories":1}},"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":0,"samples":[{"code_sha256_prefix":"0e5de1a0b68887a8","entry":"create_parser","repo":"humancompatibleai/reducing-exploitability","repo_kind":"official","path":"src/dep/evaluate.py","file_url":"https://github.com/humancompatibleai/reducing-exploitability/blob/HEAD/src/dep/evaluate.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"0e5de1a0b68887a8"}},{"code_sha256_prefix":"f73e792e0b5295b0","entry":"default_policy_agent_mapping","repo":"humancompatibleai/reducing-exploitability","repo_kind":"official","path":"src/dep/evaluate.py","file_url":"https://github.com/humancompatibleai/reducing-exploitability/blob/HEAD/src/dep/evaluate.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"f73e792e0b5295b0"}},{"code_sha256_prefix":"b7426b39719abff3","entry":"get_checkpoint_file","repo":"humancompatibleai/reducing-exploitability","repo_kind":"official","path":"src/aprl_defense/common/io.py","file_url":"https://github.com/humancompatibleai/reducing-exploitability/blob/HEAD/src/aprl_defense/common/io.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"b7426b39719abff3"}},{"code_sha256_prefix":"0744420afc4631e3","entry":"get_saved_config","repo":"humancompatibleai/reducing-exploitability","repo_kind":"official","path":"src/aprl_defense/common/io.py","file_url":"https://github.com/humancompatibleai/reducing-exploitability/blob/HEAD/src/aprl_defense/common/io.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"0744420afc4631e3"}},{"code_sha256_prefix":"21303ac8a9098ecf","entry":"save_params","repo":"humancompatibleai/reducing-exploitability","repo_kind":"official","path":"src/aprl_defense/common/rllib_io.py","file_url":"https://github.com/humancompatibleai/reducing-exploitability/blob/HEAD/src/aprl_defense/common/rllib_io.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"21303ac8a9098ecf"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}