{"id":"W4416748856","doi":"10.1109/iros60139.2025.11246761","title":"Refined Policy Distillation: From VLA Generalists to RL Experts","year":2025,"lang":"","type":"article","venue":"","topic":"Multimodal Machine Learning Applications","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"Ministry of Education","keywords":"Reinforcement learning; Complement (music); Policy learning; Generalization; Perspective (graphical); Key (lock); Task (project management)","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002816087,0.00186076,0.001530698,0.0007344913,0.0006344117,0.00151631,0.003217333,0.001864276,0.005874827],"category_scores_gemma":[0.01255986,0.001116929,0.001234031,0.0004563435,0.001709372,0.00235454,0.003758692,0.005168702,0.001872565],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001482437,"about_ca_system_score_gemma":0.002771551,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00654702,"about_ca_topic_score_gemma":0.009019329,"domain_scores_codex":[0.9985689,0.0005343268,0.0000738622,0.0004166111,0.000228283,0.0001780331],"domain_scores_gemma":[0.9958924,0.002436742,0.0002267439,0.0009094849,0.0002966843,0.0002378874],"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":[0.0004079239,0.0002385515,0.001994979,0.0002496434,0.0001051268,0.0001582778,0.0002243328,0.7833506,0.004264551,0.02006976,0.007436086,0.1815001],"study_design_scores_gemma":[0.00004197179,0.00004163349,0.0000570387,0.00001717534,0.000007829844,0.00001482602,0.00001042234,0.9904656,0.0009338133,0.007199453,0.001201069,0.000009198577],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03268072,0.0006745213,0.9514704,0.0007288114,0.0001865975,0.0002334933,0.0002872926,0.008639494,0.005098685],"genre_scores_gemma":[0.5795216,0.0003578774,0.4083356,0.001259403,0.0001152874,0.0007500713,0.001065483,0.00166888,0.006925785],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.00654702,"threshold_uncertainty_score":0.0196532,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01237008961308218,"score_gpt":0.3352070523733426,"score_spread":0.3228369627602605,"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."}}