{"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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001006322,0.0002196114,0.0002803587,0.0003290093,0.0002990942,0.0000382085,0.0003930135,0.0001706714,0.000004179595],"category_scores_gemma":[0.001862822,0.0002301025,0.00008337463,0.0008258084,0.00004011824,0.0002422494,0.00001653558,0.000356915,0.000005800728],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004738537,"about_ca_system_score_gemma":0.00003347402,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003615353,"about_ca_topic_score_gemma":0.000003680374,"domain_scores_codex":[0.9985356,0.0001258847,0.0002654293,0.0005185905,0.0001323396,0.0004222007],"domain_scores_gemma":[0.9942781,0.005057681,0.00006245179,0.0003945451,0.000105318,0.0001018584],"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":[0.00001871333,0.00001247507,0.0001188887,0.00006283162,0.00003299501,0.000003277017,0.0005231001,0.6483724,0.001007729,0.0001724117,0.000006759119,0.3496684],"study_design_scores_gemma":[0.0005854437,0.0001887995,0.002729422,0.00004570744,0.00005184824,0.00004403816,0.0001272176,0.9915492,0.002845568,0.0008696761,0.0005929464,0.0003701259],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01214244,0.00002675067,0.9848694,0.0005029268,0.0008870606,0.0002851647,0.000008897476,0.001274616,0.000002712334],"genre_scores_gemma":[0.1430408,0.00002249495,0.856556,0.00004187643,0.00005586674,0.0001688277,0.000008691548,0.00003514877,0.00007026116],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.3492983,"threshold_uncertainty_score":0.9383309,"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."}}