{"id":"W4362721716","doi":"10.1145/3591870","title":"PatchCensor: Patch Robustness Certification for Transformers via Exhaustive Testing","year":2023,"lang":"en","type":"article","venue":"ACM Transactions on Software Engineering and Methodology","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":12,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Alberta","funders":"JST-Mirai Program; National Key Research and Development Program of China; Japan Society for the Promotion of Science; Natural Sciences and Engineering Research Council of Canada; National Natural Science Foundation of China; Canadian Institute for Advanced Research","keywords":"Computer science; Robustness (evolution); Artificial intelligence; Transformer; Convolutional neural network; Computer security; Software deployment; Computer engineering; Machine learning; Real-time computing; Software engineering; Electrical engineering","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":[0.005121641,0.001965404,0.001565165,0.001844844,0.0007068875,0.001347946,0.003383116,0.002140211,0.01192328],"category_scores_gemma":[0.03294856,0.0007353997,0.001693069,0.000662706,0.002803003,0.004531494,0.003572494,0.002491381,0.003412743],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001132976,"about_ca_system_score_gemma":0.002654084,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001768415,"about_ca_topic_score_gemma":0.00249844,"domain_scores_codex":[0.9945037,0.001316194,0.0003808648,0.001085058,0.002115486,0.0005986715],"domain_scores_gemma":[0.9823601,0.0085227,0.001243422,0.005057041,0.002229683,0.0005869571],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00252742,0.0006745541,0.01551617,0.001789063,0.0005094242,0.001146512,0.0003064674,0.2617339,0.0538232,0.04427672,0.044998,0.5726986],"study_design_scores_gemma":[0.0002518859,0.0008043732,0.00152368,0.00014863,0.00008216672,0.0006318217,0.000111884,0.9323044,0.02379449,0.03302827,0.007259633,0.00005889936],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.08708649,0.002319758,0.8522493,0.001089546,0.0004934414,0.0006606051,0.001037674,0.04225977,0.01280331],"genre_scores_gemma":[0.8023955,0.000648586,0.1855823,0.0007987013,0.0001638575,0.0004488515,0.002458412,0.002786858,0.004717052],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01192328,"threshold_uncertainty_score":0.03988737,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1333034867614891,"score_gpt":0.3331364074471863,"score_spread":0.1998329206856971,"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."}}