{"id":"W3120394327","doi":"10.1007/s10515-021-00313-x","title":"Faults in deep reinforcement learning programs: a taxonomy and a detection approach","year":2021,"lang":"en","type":"article","venue":"Automated Software Engineering","topic":"Reinforcement Learning in Robotics","field":"Computer Science","cited_by":38,"is_retracted":false,"has_abstract":false,"ca_institutions":"Polytechnique Montréal","funders":"Fonds de Recherche du Québec-Société et Culture; Canadian Network for Research and Innovation in Machining Technology, Natural Sciences and Engineering Research Council of Canada","keywords":"Reinforcement learning; Computer science; Categorization; Artificial intelligence; Fault detection and isolation; Software; Machine learning; Software engineering; Programming language","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":[0.002922758,0.0009819542,0.001052786,0.002441832,0.00051834,0.002378521,0.002486604,0.003498669,0.0009522244],"category_scores_gemma":[0.01958479,0.0008892048,0.000962648,0.001937143,0.003147982,0.004881529,0.001945743,0.002899152,0.0001688917],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001508379,"about_ca_system_score_gemma":0.001113128,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001970613,"about_ca_topic_score_gemma":0.001286761,"domain_scores_codex":[0.9966654,0.0008114658,0.0004017906,0.0006194112,0.00122787,0.0002738941],"domain_scores_gemma":[0.9746957,0.01811218,0.003027591,0.001775602,0.001955396,0.0004335797],"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":[0.0004142182,0.0003330157,0.02029547,0.0008640878,0.0001290486,0.0009581085,0.000653554,0.3742052,0.00456732,0.324276,0.002746833,0.2705571],"study_design_scores_gemma":[0.00001271914,0.00008503461,0.0007017835,0.00006793698,0.00002965604,0.0002882594,0.00005012322,0.8528542,0.002245684,0.1426641,0.0009804758,0.0000199686],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03495555,0.001976957,0.9594291,0.0007904185,0.00004332014,0.00008715279,0.0000961708,0.0007437037,0.001877576],"genre_scores_gemma":[0.7887232,0.001757696,0.206377,0.0002403409,0.0001913661,0.0001690634,0.0001843849,0.0001241721,0.002232801],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.003498669,"threshold_uncertainty_score":0.01545721,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01206261324535972,"score_gpt":0.2026977882953615,"score_spread":0.1906351750500018,"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."}}