{"id":"W4226207655","doi":"10.1109/qrs54544.2021.00118","title":"Understanding the Resilience of Neural Network Ensembles against Faulty Training Data","year":2021,"lang":"en","type":"article","venue":"2021 IEEE 21st International Conference on Software Quality, Reliability and Security (QRS)","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":16,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of British Columbia","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Resilience (materials science); Machine learning; Artificial intelligence; Artificial neural network; Training set; Ensemble learning; Entropy (arrow of time); Training (meteorology); Ensemble forecasting; Data mining","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":[0.003874542,0.0007348652,0.0005928439,0.001013265,0.0005416111,0.0009133366,0.0008083643,0.001135713,0.0008754389],"category_scores_gemma":[0.03136592,0.0003775825,0.0004592625,0.0004011126,0.001211602,0.002672958,0.001504632,0.001457868,0.0001993361],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009369771,"about_ca_system_score_gemma":0.0005655292,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003112675,"about_ca_topic_score_gemma":0.002021557,"domain_scores_codex":[0.9989707,0.0003221009,0.00006913585,0.0002051442,0.0002471028,0.00018568],"domain_scores_gemma":[0.9868471,0.008194512,0.001468663,0.001933782,0.001153794,0.0004020838],"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":[0.0001035535,0.00003245485,0.00836568,0.00003908938,0.000063856,0.00007945106,0.00009855825,0.9697735,0.004004749,0.002623571,0.000332222,0.01448332],"study_design_scores_gemma":[0.000004226806,0.00009639929,0.002738444,0.00002044206,0.00002260112,0.000054777,0.00006212686,0.984557,0.003950556,0.008213253,0.0002667853,0.00001353085],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.7650633,0.0008904764,0.2284366,0.001495322,0.0001233576,0.00006528535,0.0002284361,0.0006576907,0.003039623],"genre_scores_gemma":[0.9931665,0.000127588,0.006163506,0.00008485196,0.00002403298,0.00002289597,0.00009099858,0.00002756861,0.0002919816],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.003874542,"threshold_uncertainty_score":0.02049077,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2138340083096355,"score_gpt":0.3701135125969131,"score_spread":0.1562795042872776,"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."}}