{"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002990438,0.001312328,0.001428939,0.000529943,0.0004165554,0.0011691,0.002188439,0.002250895,0.003157824],"category_scores_gemma":[0.01344838,0.000961195,0.0006783615,0.0004557518,0.001716092,0.002074993,0.004201536,0.004119831,0.0006534735],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000985819,"about_ca_system_score_gemma":0.001391925,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002350871,"about_ca_topic_score_gemma":0.001874054,"domain_scores_codex":[0.9990242,0.0002881702,0.00004636806,0.0002343265,0.0002985622,0.0001084649],"domain_scores_gemma":[0.9953253,0.002645888,0.0004901821,0.0007964603,0.000578813,0.0001634083],"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":[0.000203584,0.00003709633,0.0004678577,0.0001167949,0.00004451976,0.00005207762,0.00008697405,0.8938586,0.004382037,0.01116403,0.001359106,0.08822737],"study_design_scores_gemma":[0.000003268923,0.0000228155,0.00005631103,0.000007448721,0.000002822482,0.000009099595,0.000002416072,0.9941002,0.00074007,0.004897022,0.0001552427,0.000003315152],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.006892612,0.0001745943,0.9916223,0.0001495445,0.00002277316,0.00001418905,0.00002581997,0.000445856,0.0006523087],"genre_scores_gemma":[0.8064629,0.000245355,0.187721,0.0002662542,0.0001028552,0.0001229702,0.0001424458,0.0003258595,0.004610395],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.003157824,"threshold_uncertainty_score":0.01581514,"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."}}