{"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003317991,0.001120791,0.0009542161,0.0007682399,0.0008008051,0.001255089,0.001785333,0.001426049,0.001596069],"category_scores_gemma":[0.01752509,0.0007713069,0.0009265146,0.0004549177,0.001108725,0.002621062,0.003236003,0.002192501,0.001004763],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006951132,"about_ca_system_score_gemma":0.0007765864,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001571944,"about_ca_topic_score_gemma":0.002102071,"domain_scores_codex":[0.9983559,0.0006219305,0.00009213595,0.0003809375,0.0003680805,0.0001809055],"domain_scores_gemma":[0.9913478,0.004585331,0.0005977349,0.002263789,0.000935676,0.0002695426],"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":[0.0001131692,0.00009084454,0.005326533,0.00005816873,0.0001209075,0.0001952356,0.0002494769,0.9051473,0.009972098,0.002717111,0.001225111,0.07478399],"study_design_scores_gemma":[0.000004351283,0.00004891652,0.0003808273,0.00001374675,0.00001602392,0.00005909153,0.00004501285,0.9881335,0.00559643,0.00491995,0.0007717304,0.00001034767],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1408897,0.0002729702,0.8520386,0.0005761452,0.00009259106,0.0001044814,0.00008323487,0.003928138,0.00201423],"genre_scores_gemma":[0.8332716,0.0001573994,0.1639609,0.0002987531,0.00006058392,0.0001259699,0.0002313883,0.0003643853,0.001529024],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.003317991,"threshold_uncertainty_score":0.01754743,"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."}}