{"id":"W4401838084","doi":"10.2139/ssrn.4935952","title":"Trimming the Risk: Towards Reliable Continuous Training for Deep Learning Inspection Systems","year":2024,"lang":"en","type":"preprint","venue":"SSRN Electronic Journal","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":false,"ca_institutions":"Polytechnique Montréal","funders":"","keywords":"Trimming; Training (meteorology); Computer science; Artificial intelligence; Risk analysis (engineering); Operations management; Business; Engineering; Geography","routes":{"ca_aff":true,"ca_fund":false,"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":["metaepi_narrow","sts","scholarly_communication","research_integrity"],"consensus_categories":[],"category_scores_codex":[0.01182014,0.0005311563,0.0007114311,0.0003837732,0.001737164,0.001791317,0.002247477,0.0004201992,0.000002207956],"category_scores_gemma":[0.001183212,0.0004059824,0.0005576651,0.000465935,0.00007318232,0.0003149358,0.001342024,0.01994867,0.00001967297],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002634805,"about_ca_system_score_gemma":0.004705556,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0006525355,"about_ca_topic_score_gemma":0.0001567397,"domain_scores_codex":[0.9938391,0.0006663179,0.0008622527,0.000914119,0.0007041883,0.003013994],"domain_scores_gemma":[0.9973194,0.0004436293,0.001206322,0.0006323241,0.0002884835,0.0001098158],"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":[0.00004339225,0.00001811272,0.0001097755,0.0001044552,0.0005844613,0.00001247469,0.004449545,0.6431778,0.00001614178,0.180962,0.0000761507,0.1704457],"study_design_scores_gemma":[0.0005177236,0.0002901261,0.0000320831,0.000325463,0.0002353015,0.0007311006,0.006049518,0.7488862,0.000007527678,0.2372748,0.005195271,0.0004549233],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.008102906,0.01893235,0.9621118,0.001084394,0.007362571,0.0007233709,0.000002122061,0.0006768645,0.0010036],"genre_scores_gemma":[0.9837753,0.00323531,0.006851859,0.00002728506,0.003584214,0.0001300838,0.000008342322,0.0001209784,0.002266627],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9756724,"threshold_uncertainty_score":0.9998392,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01518606958273447,"score_gpt":0.2685198439459032,"score_spread":0.2533337743631687,"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."}}