{"id":"W3203969037","doi":"10.1145/3470006","title":"Automatic Fault Detection for Deep Learning Programs Using Graph Transformations","year":2021,"lang":"en","type":"article","venue":"ACM Transactions on Software Engineering and Methodology","topic":"Software Engineering Research","field":"Computer Science","cited_by":29,"is_retracted":false,"has_abstract":true,"ca_institutions":"Polytechnique Montréal","funders":"Natural Sciences and Engineering Research Council of Canada; Fonds de recherche du Québec; Institut de Valorisation des Données","keywords":"Computer science; Metamodeling; Graph; Fault detection and isolation; Precision and recall; Construct (python library); Artificial intelligence; Deep learning; Artificial neural network; Machine learning; Software; Process (computing); Data mining; Software engineering; Theoretical computer science; 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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0007135823,0.0002090886,0.0002780501,0.0003737511,0.0003112633,0.0001054805,0.000299211,0.0001765215,0.000007292343],"category_scores_gemma":[0.002779721,0.0002314364,0.0001408927,0.0007691299,0.00003083231,0.0003046611,0.0000157822,0.0004646184,0.000002411677],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000705425,"about_ca_system_score_gemma":0.00005367237,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001202928,"about_ca_topic_score_gemma":0.000008983628,"domain_scores_codex":[0.998506,0.0001835087,0.0002749627,0.0004145665,0.000177745,0.0004432091],"domain_scores_gemma":[0.9952903,0.003920874,0.00003741829,0.0004520511,0.0001599378,0.0001394088],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000005621251,0.00004349812,0.0000568534,0.0001930896,0.00008607315,0.000006694976,0.000779035,0.22969,0.003731286,0.0001792065,5.047074e-7,0.7652282],"study_design_scores_gemma":[0.0008128026,0.0003343276,0.001040039,0.00009883134,0.00007732527,0.0004986106,0.0001387184,0.9637273,0.03038945,0.0009361421,0.001474694,0.0004717393],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.05815283,0.000305241,0.9394355,0.00007355867,0.0005763309,0.0002601788,0.000002812501,0.001192806,6.961164e-7],"genre_scores_gemma":[0.2023429,0.00005829395,0.7973247,0.0000189344,0.00002803562,0.0001694396,0.000005160486,0.00003044264,0.00002211084],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.7647564,"threshold_uncertainty_score":0.9437702,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.08467335479524384,"score_gpt":0.3304131342028107,"score_spread":0.2457397794075669,"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."}}