{"id":"W4393103270","doi":"10.1016/j.automatica.2024.111642","title":"Stabilizing reinforcement learning control: A modular framework for optimizing over all stable behavior","year":2024,"lang":"en","type":"article","venue":"Automatica","topic":"Adaptive Dynamic Programming Control","field":"Computer Science","cited_by":7,"is_retracted":false,"has_abstract":true,"ca_institutions":"Honeywell (Canada); University of British Columbia","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Reinforcement learning; Modular design; Stability (learning theory); Reinforcement; Control (management); Computer science; Control theory (sociology); Control engineering; Engineering; Artificial intelligence; Machine learning; Structural engineering","routes":{"ca_aff":true,"ca_fund":true,"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"],"consensus_categories":[],"category_scores_codex":[0.0007851662,0.0002641902,0.0003410704,0.0001489612,0.0002036009,0.0007553005,0.0006779648,0.0001236669,0.00003432445],"category_scores_gemma":[0.00035527,0.0002498341,0.000198789,0.0002897215,0.00004985316,0.0007112919,0.000217511,0.0003710538,0.00004482116],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002232415,"about_ca_system_score_gemma":0.0001214938,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001994062,"about_ca_topic_score_gemma":0.000001265988,"domain_scores_codex":[0.9977325,0.00008123798,0.0004588138,0.0005914769,0.000419331,0.0007167084],"domain_scores_gemma":[0.9982968,0.0007620331,0.0001143353,0.000597173,0.0000901629,0.0001395186],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00002147523,0.0001034563,0.0001331205,0.0002726217,0.0003567144,0.00008870938,0.002908494,0.06795014,0.003427955,0.8457384,0.0001925341,0.07880634],"study_design_scores_gemma":[0.0004993014,0.0002078615,0.000116637,0.0002027755,0.0001005665,0.000008008171,0.00005684089,0.9829187,0.00009458934,0.009220579,0.006282474,0.0002916743],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.004116178,0.0004718861,0.9913583,0.0006349716,0.0004457641,0.001180531,0.000003381238,0.001589906,0.0001990115],"genre_scores_gemma":[0.6090257,0.000003565403,0.3900344,0.0002025531,0.00005157268,0.0004719739,0.000004677914,0.0000317072,0.000173906],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9149686,"threshold_uncertainty_score":0.9999954,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01738113249222614,"score_gpt":0.2899690171600185,"score_spread":0.2725878846677924,"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."}}