{"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0006114245,0.001315534,0.0004467904,0.001889719,0.0002707929,0.0006801109,0.001252808,0.0007514284,0.001174307],"category_scores_gemma":[0.004135099,0.0004635462,0.001465661,0.0006387085,0.0007707631,0.00152025,0.0008397371,0.001122014,0.000275026],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001406141,"about_ca_system_score_gemma":0.001089613,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006261319,"about_ca_topic_score_gemma":0.009909898,"domain_scores_codex":[0.9991436,0.0002001984,0.00006612291,0.0002630139,0.0002543865,0.00007269775],"domain_scores_gemma":[0.9977944,0.001214436,0.0003638065,0.0003476657,0.0002463458,0.00003332853],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0003166775,0.000251413,0.0151965,0.0005430505,0.000144715,0.0006474127,0.0002458072,0.5558929,0.03866959,0.01277708,0.003667376,0.3716475],"study_design_scores_gemma":[0.00001015695,0.00003416989,0.0004710676,0.00001705417,0.00001726598,0.00005333234,0.00002266652,0.9765786,0.01274974,0.009288052,0.0007522127,0.000005631972],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.09719545,0.0003259618,0.8797896,0.0002979476,0.00002886005,0.000123898,0.0006614685,0.02060179,0.0009750662],"genre_scores_gemma":[0.6651152,0.0002234419,0.3307129,0.000163537,0.0000138852,0.0001793158,0.001963964,0.0006599888,0.0009677524],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.006261319,"threshold_uncertainty_score":0.01244974,"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."}}