{"id":"W7124294995","doi":"10.65109/igaq8141","title":"Taming Multi-Agent Reinforcement Learning with Estimator Variance Reduction","year":2025,"lang":"","type":"article","venue":"","topic":"Reinforcement Learning in Robotics","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Alberta","funders":"","keywords":"Estimator; Variance (accounting); Reinforcement learning; Variance reduction; Range (aeronautics); Sample (material); Reduction (mathematics)","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","sts","scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.001011016,0.0006908292,0.0005656595,0.0005025297,0.001393266,0.00118893,0.001342342,0.0002396056,0.0002756898],"category_scores_gemma":[0.0002554859,0.000647655,0.0001526615,0.001888572,0.0002613445,0.00134006,0.0009711587,0.001111469,0.000335434],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006954734,"about_ca_system_score_gemma":0.0007751693,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001394999,"about_ca_topic_score_gemma":0.000001448268,"domain_scores_codex":[0.9950272,0.0002481792,0.001168419,0.001359835,0.001036585,0.001159826],"domain_scores_gemma":[0.9970803,0.0001426651,0.0006807234,0.001375717,0.0004630679,0.0002574777],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00005011758,0.00007348459,0.0005809635,0.000214872,0.0002225681,0.00001900589,0.001591058,0.9411837,0.0002801263,0.04094274,0.0003626197,0.01447874],"study_design_scores_gemma":[0.001759423,0.0005683071,0.000408356,0.001017606,0.0001219777,0.00003030092,0.0006747048,0.9796378,0.002009824,0.00001714926,0.01306493,0.0006896283],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0002792134,0.000209867,0.9669511,0.001149,0.002614105,0.0009630133,1.624435e-7,0.0005045199,0.02732905],"genre_scores_gemma":[0.553359,0.0001284477,0.290018,0.0002553572,0.00009321443,0.00005009634,0.00000675733,0.00003467573,0.1560545],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.6769331,"threshold_uncertainty_score":0.9999068,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01878255743847185,"score_gpt":0.2685703962063445,"score_spread":0.2497878387678727,"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."}}