{"id":"W4413144748","doi":"10.1109/cvpr52734.2025.01367","title":"Can Large Vision-Language Models Correct Semantic Grounding Errors By Themselves?","year":2025,"lang":"en","type":"article","venue":"","topic":"Multimodal Machine Learning Applications","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Vector Institute","funders":"","keywords":"Computer science; Natural language processing; Ground; Artificial intelligence; Engineering","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":[0.0003400405,0.000161064,0.0001672853,0.0001583169,0.0002998438,0.000204171,0.001028459,0.00006087924,0.00004394796],"category_scores_gemma":[0.00005637814,0.0001410969,0.00006376022,0.0007001575,0.00002530458,0.0002904254,0.0003912313,0.0002410653,0.000066849],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00007058264,"about_ca_system_score_gemma":0.00005855365,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002503071,"about_ca_topic_score_gemma":0.0001716436,"domain_scores_codex":[0.9986342,0.00008249143,0.0002147429,0.0004954975,0.0002163281,0.0003567828],"domain_scores_gemma":[0.9988048,0.0002013986,0.00006799857,0.0008000702,0.00004060593,0.0000851048],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000004761147,0.0003597439,0.004443881,0.00005176106,0.00006217841,0.000008335513,0.005305467,0.004982634,0.008209815,0.9083492,0.02533419,0.04288799],"study_design_scores_gemma":[0.0002543346,0.00001665944,0.001830354,0.00002949512,0.000006899262,0.000003392443,0.0001993596,0.9892314,0.0005420286,0.006881724,0.0008300184,0.0001743544],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.08114185,0.0001009346,0.8916423,0.004052828,0.000213877,0.0002239944,0.000005475587,0.0005400685,0.02207869],"genre_scores_gemma":[0.976265,0.000004094677,0.01713943,0.001155122,0.00001589196,0.00003958953,0.00001289593,0.00001141634,0.00535662],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9842488,"threshold_uncertainty_score":0.5753763,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.007488223827787919,"score_gpt":0.2949452169845906,"score_spread":0.2874569931568027,"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."}}