{"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":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.003151857,0.0002950553,0.0004578246,0.00006685985,0.0004862452,0.0004328362,0.00259025,0.0001645535,0.00009706607],"category_scores_gemma":[0.003878123,0.0002543586,0.0001218314,0.0004611015,0.0004890763,0.0008398221,0.001558398,0.0008652725,0.000005357891],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001411155,"about_ca_system_score_gemma":0.0004195886,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001323403,"about_ca_topic_score_gemma":0.0002857566,"domain_scores_codex":[0.9953601,0.00118734,0.000793685,0.001157027,0.00108271,0.0004191853],"domain_scores_gemma":[0.9949145,0.00227259,0.0004337038,0.001695772,0.0005479175,0.0001354591],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001119756,0.0002269415,0.01488284,0.0001496199,0.0001221465,0.00004566487,0.009562735,0.06821425,0.000194248,0.8874332,0.0004943064,0.0185621],"study_design_scores_gemma":[0.0008073011,0.0001036216,0.005046624,0.0003962851,0.00002955314,0.00003292654,0.007069936,0.7965906,0.000173182,0.1878888,0.001214547,0.0006465896],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1580515,0.0001380285,0.8268261,0.007924349,0.002494024,0.0002800206,0.0002078574,0.000150631,0.003927496],"genre_scores_gemma":[0.987369,0.0002158213,0.01127161,0.0006608759,0.0002818757,0.000007805702,0.0001053033,0.00001278379,0.00007494519],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8293175,"threshold_uncertainty_score":0.9999909,"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."}}