{"id":"W4416177363","doi":"10.48550/arxiv.2508.06836","title":"Multi-level Advantage Credit Assignment for Cooperative Multi-Agent Reinforcement Learning","year":2025,"lang":"en","type":"preprint","venue":"ArXiv.org","topic":"Reinforcement Learning in Robotics","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Alliance de recherche numérique du Canada","keywords":"Counterfactual thinking; Reinforcement learning; Key (lock); Diversity (politics); Function (biology)","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":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.001005797,0.0009352987,0.0008971014,0.0003676224,0.0006680608,0.000449695,0.002743653,0.0005143566,0.00007323432],"category_scores_gemma":[0.0008364983,0.0009378311,0.0004500104,0.0003296037,0.0001228003,0.0004448374,0.005047799,0.001759157,0.0002175449],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009294432,"about_ca_system_score_gemma":0.0007517567,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0000490329,"about_ca_topic_score_gemma":0.000009524138,"domain_scores_codex":[0.9947892,0.0002723635,0.001287377,0.001733097,0.0008366349,0.001081331],"domain_scores_gemma":[0.9958959,0.0004553596,0.0009243965,0.001822512,0.000615472,0.000286424],"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":[0.00002468214,0.0001257707,0.009800163,0.0003239539,0.0003254193,0.00002024356,0.001555847,0.9840989,0.0005878258,0.0007364635,0.000953489,0.001447238],"study_design_scores_gemma":[0.002245718,0.0003558267,0.006223835,0.0006051159,0.0001019171,0.00000226975,0.0002519187,0.9638681,0.003153569,0.00001003476,0.02217247,0.001009264],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.002121607,0.0002024877,0.9891922,0.0005149751,0.003785552,0.002755241,0.00003268824,0.000547563,0.0008476132],"genre_scores_gemma":[0.47858,0.0004199655,0.4065465,0.001193953,0.0004502881,0.001491006,0.0005198748,0.0001060968,0.1106923],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.5826458,"threshold_uncertainty_score":0.9993072,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1193564411615443,"score_gpt":0.3388184746199345,"score_spread":0.2194620334583902,"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."}}