{"id":"W4388090033","doi":"10.1016/j.datak.2023.102240","title":"Hierarchical framework for interpretable and specialized deep reinforcement learning-based predictive maintenance","year":2023,"lang":"en","type":"article","venue":"Data & Knowledge Engineering","topic":"Reliability and Maintenance Optimization","field":"Engineering","cited_by":22,"is_retracted":false,"has_abstract":false,"ca_institutions":"","funders":"H2020 Marie Skłodowska-Curie Actions; Canadian Institute of Steel Construction; Horizon 2020 Framework Programme; Österreichische Forschungsförderungsgesellschaft; Science Foundation Ireland; Horizon 2020","keywords":"Interpretability; Reinforcement learning; Machine learning; Computer science; Artificial intelligence; Probabilistic logic; Black box; Context (archaeology); Markov decision process; Risk analysis (engineering); Markov process","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0006604235,0.0007725386,0.0009458678,0.0005279849,0.0003220009,0.0008779212,0.002520048,0.001169335,0.004899127],"category_scores_gemma":[0.002049467,0.0004596637,0.0006757112,0.0004451576,0.0005899517,0.001104024,0.001152101,0.001676633,0.000789146],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001137878,"about_ca_system_score_gemma":0.001672851,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01075487,"about_ca_topic_score_gemma":0.01719184,"domain_scores_codex":[0.9996811,0.00005381981,0.00001721659,0.0001040598,0.00008196046,0.00006183515],"domain_scores_gemma":[0.9994408,0.0002229378,0.00005810237,0.00008880058,0.0001473052,0.00004193994],"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":[0.0001122182,0.00009750725,0.0006341479,0.0000809222,0.00004457537,0.00008462937,0.00005986927,0.8639094,0.003155795,0.02021226,0.003343307,0.1082653],"study_design_scores_gemma":[0.000003140622,0.000008387973,0.00004199818,0.000002959861,0.000003545015,0.000004313174,0.000001722086,0.9956185,0.0002360683,0.003887188,0.0001900882,0.000001976793],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01030042,0.0003218107,0.9854949,0.000167126,0.00004230088,0.00003097314,0.0002245102,0.001593722,0.001824295],"genre_scores_gemma":[0.7552218,0.0002959949,0.2379251,0.0002112266,0.00008643461,0.0001721658,0.0007470979,0.0002407521,0.005099514],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01075487,"threshold_uncertainty_score":0.02138454,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0151796263670819,"score_gpt":0.2519409794108211,"score_spread":0.2367613530437392,"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."}}