{"id":"W4407450044","doi":"10.1109/tcyb.2025.3536165","title":"Personalizing Vision-Language Models With Hybrid Prompts for Zero-Shot Anomaly Detection","year":2025,"lang":"en","type":"article","venue":"IEEE Transactions on Cybernetics","topic":"Anomaly Detection Techniques and Applications","field":"Computer Science","cited_by":36,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto","funders":"South University of Science and Technology of China; Ministry of Industry and Information Technology of the People's Republic of China; China Scholarship Council","keywords":"Zero (linguistics); Anomaly detection; Shot (pellet); Computer science; Ground zero; Anomaly (physics); Artificial intelligence; Natural language processing; Physics; Linguistics; Chemistry","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":[0.0001221665,0.0002022134,0.0001672704,0.0002516815,0.0004422691,0.000157891,0.0003927219,0.00008250133,0.000007698683],"category_scores_gemma":[0.000001924159,0.0001900619,0.000124676,0.0005352168,0.0000610028,0.0003005204,0.000004108255,0.0002082986,0.000008747302],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001060239,"about_ca_system_score_gemma":0.00007723681,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00006321035,"about_ca_topic_score_gemma":0.00007098652,"domain_scores_codex":[0.998756,0.00002765386,0.0002439757,0.0005058176,0.0002010525,0.0002655373],"domain_scores_gemma":[0.9990581,0.00009885048,0.00008020052,0.0005428147,0.0001447403,0.00007529859],"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":[0.0002666824,0.001160432,0.000006938528,0.0001658586,0.0002504717,0.00001187264,0.001661255,0.08892287,0.05441764,0.03761174,0.001082773,0.8144414],"study_design_scores_gemma":[0.0005637226,0.0005709448,0.0000307952,0.00007530749,0.00005740234,0.00003186537,0.0000719945,0.5818834,0.4083477,0.005537622,0.002524133,0.0003050537],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01254366,0.00004633534,0.9841335,0.0003665655,0.0001794182,0.0007632145,0.00002244028,0.0005237933,0.001421106],"genre_scores_gemma":[0.9221298,0.00002301531,0.07482333,0.0002501847,0.00001891044,0.0005147648,0.000001764174,0.00002031073,0.002217879],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9095862,"threshold_uncertainty_score":0.7750499,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01531708000535554,"score_gpt":0.2699115341578634,"score_spread":0.2545944541525079,"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."}}