{"id":"W4408355496","doi":"10.1109/icassp49660.2025.10889927","title":"CA-UAP: Content-Agnostic Universal Adversarial Perturbation for Enhanced Generalization","year":2025,"lang":"en","type":"article","venue":"","topic":"Image Processing Techniques and Applications","field":"Engineering","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of British Columbia","funders":"Natural Science Foundation of Guangdong Province","keywords":"Adversarial system; Computer science; Generalization; Theoretical computer science; Artificial intelligence; Mathematics; Mathematical analysis","routes":{"ca_aff":true,"ca_fund":false,"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":[0.00003653054,0.00006575377,0.00006231166,0.0000544744,0.00007104911,0.00002712578,0.00006691462,0.00004945268,0.00003199761],"category_scores_gemma":[0.00003299379,0.00006726477,0.00002712177,0.0001240991,0.000009500165,0.00009713998,0.000009309841,0.00003042087,0.000004476055],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000633804,"about_ca_system_score_gemma":0.00001965891,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001380743,"about_ca_topic_score_gemma":0.000007132171,"domain_scores_codex":[0.9996746,0.000004176667,0.0001012269,0.00009336225,0.00003401727,0.00009258892],"domain_scores_gemma":[0.999765,0.00003475305,0.00001288565,0.00008618286,0.00008637783,0.0000147647],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0000297249,0.00005584024,0.00003055661,0.0002469294,0.00006755272,3.227457e-7,0.0002268435,0.02216732,0.6069282,0.2628199,0.07156349,0.03586329],"study_design_scores_gemma":[0.0006589628,0.00001936023,0.00007531008,0.00003832548,0.00004558077,2.851837e-7,0.00006193495,0.5749845,0.3841779,0.006400384,0.03336627,0.000171162],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.002217164,0.0000501928,0.9839492,0.0002533792,0.0001578277,0.0002913461,0.000008544551,0.0005022847,0.01257005],"genre_scores_gemma":[0.9521459,0.00003277581,0.03972908,0.0001726088,0.00006239241,0.0001306357,0.0001109178,0.00001274689,0.007602999],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9499287,"threshold_uncertainty_score":0.2742978,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01316616128634957,"score_gpt":0.2432701833507717,"score_spread":0.2301040220644221,"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."}}