{"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/the-mirage-of-action-dependent-baselines-in","title":"The Mirage of Action-Dependent Baselines in Reinforcement Learning","arxiv_id":"1802.10031","date":"2018-02-27","proceeding":"ICML 2018 7","authors":["George Tucker","Surya Bhupatiraju","Shixiang Gu","Richard E. Turner","Zoubin Ghahramani","Sergey Levine"],"abstract":"Policy gradient methods are a widely used class of model-free reinforcement\nlearning algorithms where a state-dependent baseline is used to reduce gradient\nestimator variance. Several recent papers extend the baseline to depend on both\nthe state and action and suggest that this significantly reduces variance and\nimproves sample efficiency without introducing bias into the gradient\nestimates. To better understand this development, we decompose the variance of\nthe policy gradient estimator and numerically show that learned\nstate-action-dependent baselines do not in fact reduce variance over a\nstate-dependent baseline in commonly tested benchmark domains. We confirm this\nunexpected result by reviewing the open-source code accompanying these prior\npapers, and show that subtle implementation decisions cause deviations from the\nmethods presented in the papers and explain the source of the previously\nobserved empirical gains. Furthermore, the variance decomposition highlights\nareas for improvement, which we demonstrate by illustrating a simple change to\nthe typical value function parameterization that can significantly improve\nperformance.","url_abs":"http://arxiv.org/abs/1802.10031v3","url_pdf":"http://arxiv.org/pdf/1802.10031v3.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":"the-mirage-of-action-dependent-baselines-in","repo_url":"https://github.com/brain-research/mirage-rl","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"policy-gradient-methods","task_name":"Policy Gradient Methods"},{"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":{"syntology_url":"https://syntology.ai/paper/1802.10031","atlas_url":"https://app.syntology.ai/?focus=1802.10031","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"1802.10031"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-25T09:33:49+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/brain-research/mirage-rl","reach":null}],"summary":{"ran_violates":1,"ran_fixture":2},"by_repo_kind":{"listed":{"samples":3,"ran":3,"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":"50ff603751d44853","entry":"is_sym","repo":"brain-research/mirage-rl","repo_kind":"listed","path":"lqg/lqr.py","file_url":"https://github.com/brain-research/mirage-rl/blob/HEAD/lqg/lqr.py","link_basis":"first_harvest_node","language":"python","status":"ran_violates","verification_level":1,"contract_check":"VIOLATES","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"50ff603751d44853"}},{"code_sha256_prefix":"b6aa5754504b157d","entry":"logdet","repo":"brain-research/mirage-rl","repo_kind":"listed","path":"lqg/lqr.py","file_url":"https://github.com/brain-research/mirage-rl/blob/HEAD/lqg/lqr.py","link_basis":"first_harvest_node","language":"python","status":"ran_fixture","verification_level":1,"contract_check":"RAISES","metamorphic_tier":"invariant","behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"b6aa5754504b157d"}},{"code_sha256_prefix":"bacd297c9286a46c","entry":"logsumdiag","repo":"brain-research/mirage-rl","repo_kind":"listed","path":"lqg/lqr.py","file_url":"https://github.com/brain-research/mirage-rl/blob/HEAD/lqg/lqr.py","link_basis":"first_harvest_node","language":"python","status":"ran_fixture","verification_level":1,"contract_check":"RAISES","metamorphic_tier":"invariant","behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"bacd297c9286a46c"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}