{"id":"W4388212577","doi":"10.1109/issrew60843.2023.00032","title":"Building Resilient ML Applications using Ensembles against Faulty Training Data","year":2023,"lang":"en","type":"article","venue":"","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of British Columbia Hospital; University of British Columbia","funders":"","keywords":"Computer science; Resilience (materials science); Reuse; Process (computing); Training set; Training (meteorology); Machine learning; Artificial intelligence; Data modeling; Software engineering; Engineering","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":[],"consensus_categories":[],"category_scores_codex":[0.0009015487,0.0001573274,0.0001709779,0.0002217294,0.0005111308,0.0002190763,0.002435512,0.00006312961,0.000008533171],"category_scores_gemma":[0.0002431696,0.0001547167,0.00003849943,0.001370303,0.00004718093,0.0008009599,0.00229902,0.0002394893,0.0000770583],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00006011895,"about_ca_system_score_gemma":0.0001200495,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00004990297,"about_ca_topic_score_gemma":0.0000106965,"domain_scores_codex":[0.9980462,0.00009085386,0.0002861755,0.000766165,0.0003664786,0.0004441762],"domain_scores_gemma":[0.997776,0.0002987068,0.0001118756,0.001658709,0.00005095608,0.0001037485],"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":[0.000002381043,0.00002166431,0.0003871507,0.00002272188,0.00002920639,0.00002900668,0.001176815,0.5552552,0.01031779,0.1378757,0.00108312,0.2937993],"study_design_scores_gemma":[0.0001371611,0.000004983669,0.0002441394,0.00002291312,0.000006058744,0.000007447926,0.0002938242,0.9813579,0.0002169995,0.001867948,0.01564237,0.0001982383],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01584667,0.00002500886,0.9791814,0.0005926535,0.0002425027,0.0002085829,0.000005927364,0.001063221,0.002834005],"genre_scores_gemma":[0.384275,0.000009115058,0.6150461,0.0001984165,0.0002027995,0.00001511452,0.00003543012,0.00002114673,0.0001969323],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.4261028,"threshold_uncertainty_score":0.6309162,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1416582844413375,"score_gpt":0.3757090690191169,"score_spread":0.2340507845777794,"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."}}