{"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":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.001157293,0.0003916501,0.0005937864,0.0003559628,0.0005921015,0.0008021506,0.003124515,0.0001895652,0.00004540989],"category_scores_gemma":[0.0008923962,0.0003581261,0.0004487422,0.0015714,0.00006469929,0.002106995,0.0008274867,0.0005225157,0.00006267698],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00008326206,"about_ca_system_score_gemma":0.0001424574,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000147417,"about_ca_topic_score_gemma":0.00006351864,"domain_scores_codex":[0.9968764,0.0001677047,0.0005178867,0.001007101,0.0005611373,0.0008698086],"domain_scores_gemma":[0.9973407,0.001057712,0.0002873159,0.0009322893,0.0002042389,0.0001777683],"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":[0.0002334048,0.0001348131,0.00141563,0.000207068,0.001455102,0.00009866499,0.002520346,0.8365723,0.0448997,0.003417508,0.003722296,0.1053232],"study_design_scores_gemma":[0.00172267,0.00008282119,0.0008045063,0.00001472533,0.0002376189,0.00001495642,0.00006677324,0.9897034,0.003279251,0.001055101,0.00253055,0.0004876349],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03361709,0.00002712491,0.9588175,0.002149342,0.002946669,0.0006136203,0.00001020487,0.001149967,0.0006685112],"genre_scores_gemma":[0.9579814,0.000004352801,0.03902736,0.0007716226,0.0008031001,0.0002561136,0.0000124861,0.00008433132,0.001059228],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9243643,"threshold_uncertainty_score":0.999887,"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."}}