{"id":"W4405033500","doi":"10.1109/iccv51701.2025.00210","title":"Visual Modality Prompt for Adapting Vision-Language Object Detectors","year":2025,"lang":"en","type":"preprint","venue":"","topic":"Advanced Image and Video Retrieval Techniques","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"École de Technologie Supérieure","funders":"Alliance de recherche numérique du Canada; Natural Sciences and Engineering Research Council of Canada; Mitacs","keywords":"Modality (human–computer interaction); Computer science; Object (grammar); Artificial intelligence; Computer vision; Psychology","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":[0.001018965,0.001527713,0.0006397414,0.0004740309,0.0002722748,0.0008618819,0.001742293,0.00116812,0.003905527],"category_scores_gemma":[0.004186337,0.0004126179,0.0007173803,0.0004072962,0.0007347427,0.001618056,0.001872265,0.001935775,0.002186291],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006010839,"about_ca_system_score_gemma":0.000933316,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003132528,"about_ca_topic_score_gemma":0.004585605,"domain_scores_codex":[0.9994379,0.0001058233,0.00002188382,0.0002455642,0.0001222861,0.00006655874],"domain_scores_gemma":[0.9993699,0.0001935571,0.00003980013,0.0001883565,0.0001701537,0.00003826194],"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":[0.0004881664,0.0002678514,0.001452192,0.0003051267,0.0001059873,0.0001702013,0.0002387082,0.0738185,0.1221212,0.009072619,0.01338705,0.7785724],"study_design_scores_gemma":[0.00007542173,0.0002139101,0.0008598178,0.00003672163,0.00006394619,0.0002058527,0.00008011171,0.8892694,0.0842358,0.01548644,0.009423168,0.00004936273],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01501976,0.000374878,0.9708147,0.0001440142,0.0001502785,0.0001074982,0.0002661258,0.01096803,0.002154653],"genre_scores_gemma":[0.3682014,0.0003160592,0.6211778,0.0007370947,0.0001147746,0.0002938299,0.001229613,0.001406851,0.006522539],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.003905527,"threshold_uncertainty_score":0.01306534,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02019209462687104,"score_gpt":0.3795878834902924,"score_spread":0.3593957888634214,"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."}}