{"id":"W4287865816","doi":"10.1115/1.4068705","title":"The Search for Feedback in Reinforcement Learning","year":2025,"lang":"en","type":"preprint","venue":"Journal of Dynamic Systems Measurement and Control","topic":"Neural Networks and Reservoir Computing","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Rockwell Automation (Canada)","funders":"","keywords":"Reinforcement learning; Parametrization (atmospheric modeling); Representation (politics); Computer science; Nonlinear system; Sequence (biology); State (computer science); Feedback loop; Open-loop controller; Control theory (sociology); Dynamical systems theory; Mathematical optimization; Closed loop; Artificial intelligence; Mathematics; Algorithm; Law; Control engineering; Engineering; Control (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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.006660847,0.0002123173,0.0005418531,0.0001890706,0.0002708174,0.0006066619,0.001015331,0.0001362728,1.441532e-7],"category_scores_gemma":[0.0001125611,0.0001324956,0.0001797708,0.0001279227,0.00002529283,0.00009658308,0.0003855306,0.0009414774,2.076253e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003018387,"about_ca_system_score_gemma":0.0003675509,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003858747,"about_ca_topic_score_gemma":0.00003202329,"domain_scores_codex":[0.9970575,0.0004155341,0.001024802,0.0002558295,0.0008735226,0.0003728185],"domain_scores_gemma":[0.9976552,0.000439791,0.0007635382,0.0002904173,0.0007701081,0.00008090763],"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":[0.0001469043,0.0000274899,0.0009111533,0.0005608237,0.0003573072,0.00001431706,0.0002862814,0.951135,0.0002280658,0.00159437,0.0002432594,0.04449498],"study_design_scores_gemma":[0.00169738,0.0001999789,0.0006362601,0.002026548,0.00003296639,0.00001648977,0.0001261927,0.9934638,0.000005684557,0.0002291077,0.001430125,0.0001355401],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.007043651,0.02089827,0.9634568,0.00274837,0.004355101,0.001275374,8.734803e-7,0.00001806751,0.0002034771],"genre_scores_gemma":[0.9984348,0.0005476631,0.0002662037,0.00003747693,0.000258375,0.00003127658,3.564948e-7,0.000006811482,0.000416978],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9913912,"threshold_uncertainty_score":0.5850055,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02384696423610495,"score_gpt":0.258297796052063,"score_spread":0.234450831815958,"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."}}