{"id":"W4386647553","doi":"10.1016/j.compchemeng.2023.108413","title":"Control invariant set enhanced safe reinforcement learning: Improved sampling efficiency, guaranteed stability and robustness","year":2023,"lang":"en","type":"article","venue":"Computers & Chemical Engineering","topic":"Advanced Control Systems Optimization","field":"Engineering","cited_by":9,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Alberta","funders":"Natural Sciences and Engineering Research Council of Canada; Alberta Innovates","keywords":"Robustness (evolution); Reinforcement learning; Control theory (sociology); Invariant (physics); Mathematics; Stability (learning theory); Computer science; Mathematical optimization; Engineering; Artificial intelligence; Control (management); Machine learning; Chemistry","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001356626,0.0006065762,0.0007861152,0.0003035858,0.0002435703,0.0006508388,0.0009891085,0.0006450542,0.001548743],"category_scores_gemma":[0.004590947,0.0002764564,0.0003605703,0.0002149218,0.000900608,0.0006808368,0.001183545,0.001280542,0.0002258613],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005283537,"about_ca_system_score_gemma":0.00110614,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002692459,"about_ca_topic_score_gemma":0.002049172,"domain_scores_codex":[0.9994281,0.0001433393,0.00002516623,0.00009531988,0.000212623,0.00009536853],"domain_scores_gemma":[0.9980975,0.0009691403,0.0002083144,0.0002313449,0.0003784153,0.0001153999],"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":[0.0003834676,0.0001475114,0.0008091482,0.00007697885,0.00004992684,0.0001145467,0.00007737689,0.9068632,0.008830855,0.01817373,0.0007037877,0.0637694],"study_design_scores_gemma":[0.00001407294,0.00004524945,0.00004681364,0.000001817845,0.000003059105,0.000009211456,0.000001726915,0.9971257,0.0008638735,0.0018149,0.00007140992,0.000002141712],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.05798344,0.0001243006,0.939137,0.0001132372,0.00003766199,0.00003175409,0.00002210393,0.00040567,0.002144766],"genre_scores_gemma":[0.9573963,0.00004775779,0.0409614,0.00005490957,0.00001586404,0.00004238715,0.0000358047,0.00004576701,0.001399854],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.002692459,"threshold_uncertainty_score":0.007174611,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.00995526697015245,"score_gpt":0.2032192585718763,"score_spread":0.1932639916017238,"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."}}