{"id":"W4389352648","doi":"10.1109/access.2023.3339542","title":"A Novel Semi-Supervised Adversarially Learned Meta-Classifier for Detecting Neural Trojan Attacks","year":2023,"lang":"en","type":"article","venue":"IEEE Access","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"York University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; MNIST database; Artificial intelligence; Trojan; Classifier (UML); Artificial neural network; Machine learning; Inference; Pattern recognition (psychology); Adversary; Training set; Deep learning; Computer security","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":[0.001189973,0.001321545,0.001272859,0.0008285345,0.0003374123,0.0006678575,0.002260714,0.001426955,0.0009758617],"category_scores_gemma":[0.003127556,0.0004196805,0.0009529001,0.0002938255,0.0006976008,0.001553283,0.001117671,0.001949003,0.0005479241],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008161137,"about_ca_system_score_gemma":0.0008976734,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001667071,"about_ca_topic_score_gemma":0.002806376,"domain_scores_codex":[0.9991448,0.0001992416,0.00004149022,0.0002215497,0.0002719715,0.0001210094],"domain_scores_gemma":[0.9983272,0.0005601146,0.0002906374,0.0003116526,0.0004067885,0.0001035784],"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":[0.0003720469,0.0003073323,0.006369901,0.0001276664,0.0002514294,0.0003343727,0.00007657484,0.7084273,0.01723735,0.004574536,0.006422837,0.2554987],"study_design_scores_gemma":[0.000004004966,0.00005217683,0.0001687647,0.000005026146,0.00001038739,0.0000478785,0.000003099036,0.9967021,0.001921522,0.0008223442,0.0002577282,0.000004973344],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1006418,0.001565417,0.8885686,0.0005525527,0.0002099096,0.0001486165,0.0002997874,0.004709669,0.003303668],"genre_scores_gemma":[0.9019481,0.0003077834,0.09149709,0.0005458946,0.00008779568,0.0001150484,0.0006996718,0.0001230529,0.004675546],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.002260714,"threshold_uncertainty_score":0.006293297,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1723322725974216,"score_gpt":0.3743351698200402,"score_spread":0.2020028972226186,"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."}}