{"id":"W4296182512","doi":"10.31234/osf.io/z8yrv","title":"Action chunking as conditional policy compression","year":2022,"lang":"en","type":"preprint","venue":"","topic":"Reinforcement Learning in Robotics","field":"Computer Science","cited_by":10,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"Office of Naval Research; National Science Foundation","keywords":"Chunking (psychology); Action (physics); Compression (physics); Computer science; Econometrics; Natural language processing; Artificial intelligence; Economics","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":[0.0008971201,0.0006576179,0.0006419026,0.000520608,0.0003392936,0.001070771,0.001129204,0.0008363963,0.004463222],"category_scores_gemma":[0.01024458,0.000445453,0.0004946445,0.0004703393,0.001757474,0.003330752,0.00135325,0.001447769,0.0003817476],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001273699,"about_ca_system_score_gemma":0.001216186,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003291991,"about_ca_topic_score_gemma":0.001999743,"domain_scores_codex":[0.99927,0.0001798067,0.00005021612,0.000219882,0.0001818256,0.00009819516],"domain_scores_gemma":[0.9932707,0.004029379,0.000861812,0.0011174,0.0004309578,0.000289784],"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":[0.0006384457,0.0002807899,0.004736216,0.0003448473,0.0001544352,0.000517745,0.0006553899,0.6408086,0.02563858,0.1955006,0.002600694,0.1281236],"study_design_scores_gemma":[0.00003057272,0.00009308829,0.001698295,0.00003086161,0.00003316545,0.0001092916,0.00003332753,0.840206,0.005411332,0.1507248,0.001601361,0.00002787651],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1633581,0.0006591121,0.8223906,0.001244175,0.0001271268,0.0001512609,0.0003897267,0.001171636,0.01050839],"genre_scores_gemma":[0.9175553,0.0003157295,0.07779314,0.0002348956,0.00007101373,0.000194022,0.0002317852,0.0001267863,0.003477303],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004463222,"threshold_uncertainty_score":0.01493096,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05507145678160796,"score_gpt":0.3614701003934595,"score_spread":0.3063986436118515,"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."}}