{"id":"W4405767400","doi":"10.48550/arxiv.2412.17123","title":"Fairness in Reinforcement Learning with Bisimulation Metrics","year":2024,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Ethics and Social Impacts of AI","field":"Social Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Natural Sciences and Engineering Research Council of Canada; Alliance de recherche numérique du Canada","keywords":"Bisimulation; Reinforcement learning; Reinforcement; Computer science; Psychology; Cognitive psychology; Artificial intelligence; Mathematical economics; Mathematics; Social psychology; Theoretical computer science","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01528839,0.001459784,0.001563122,0.00129106,0.0009543875,0.002252458,0.001565286,0.001523883,0.001815971],"category_scores_gemma":[0.06207263,0.0005648026,0.0007468726,0.0008519637,0.003687918,0.004605894,0.004057518,0.00293238,0.0002568115],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00365495,"about_ca_system_score_gemma":0.003112366,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002717631,"about_ca_topic_score_gemma":0.001516575,"domain_scores_codex":[0.9899621,0.006390541,0.0004804342,0.001230452,0.00142405,0.0005124143],"domain_scores_gemma":[0.9649719,0.02628626,0.002745162,0.002288486,0.002324171,0.001384049],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0002781808,0.0001532364,0.003046631,0.0001285253,0.00009524896,0.0000560443,0.0003410249,0.6345791,0.001385388,0.3172514,0.0005604411,0.04212477],"study_design_scores_gemma":[0.00002753207,0.0001029639,0.0002161275,0.00001845715,0.00001024076,0.00001172386,0.00002275681,0.8036231,0.000629771,0.19483,0.0004930154,0.0000143256],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.04759642,0.0003437927,0.9484075,0.0005484567,0.00004787839,0.00009517207,0.00005529057,0.0001835779,0.002722006],"genre_scores_gemma":[0.8655527,0.0002261606,0.1318402,0.0001847444,0.00006912537,0.0003696241,0.00008711532,0.00009690956,0.00157342],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01528839,"threshold_uncertainty_score":0.08085376,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1293072213130617,"score_gpt":0.2667864762248058,"score_spread":0.1374792549117441,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}