{"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":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0007303079,0.0004087083,0.0005073593,0.0002422219,0.0001793436,0.0003440806,0.001621995,0.0003187873,0.00001278971],"category_scores_gemma":[0.0004894412,0.0003591518,0.0003285785,0.0003490219,0.00004572484,0.0004263904,0.003258125,0.0005574157,0.000004618355],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001478956,"about_ca_system_score_gemma":0.0003327844,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001055805,"about_ca_topic_score_gemma":0.00001664371,"domain_scores_codex":[0.9974112,0.00009928124,0.0005181395,0.001160545,0.0003445629,0.0004662593],"domain_scores_gemma":[0.9978466,0.0004093382,0.0002558576,0.001127272,0.0002588496,0.0001020498],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.00006327743,0.0002167328,0.00008654204,0.001043601,0.0001081875,0.00002423666,0.0009758354,0.0002363047,0.004803714,0.011845,0.001419639,0.9791769],"study_design_scores_gemma":[0.0005365177,0.0006634594,0.0002621421,0.0008152888,0.00004714013,0.000006521971,0.0001158477,0.3403258,0.5683913,0.08473252,0.002760329,0.00134307],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.002468793,0.0003775831,0.990717,0.000271813,0.0004755327,0.001526369,0.00003180864,0.001668039,0.002463048],"genre_scores_gemma":[0.2669123,0.00004016073,0.7296288,0.0003680928,0.0001738538,0.0003293627,0.00002380179,0.00002289471,0.002500685],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.9778339,"threshold_uncertainty_score":0.999886,"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."}}