{"id":"W4402727410","doi":"10.1109/cvpr52733.2024.01307","title":"Contrasting Intra-Modal and Ranking Cross-Modal Hard Negatives to Enhance Visio-Linguistic Compositional Understanding","year":2024,"lang":"en","type":"article","venue":"","topic":"Multimodal Machine Learning Applications","field":"Computer Science","cited_by":8,"is_retracted":false,"has_abstract":true,"ca_institutions":"Mila - Quebec Artificial Intelligence Institute","funders":"","keywords":"Modal; Ranking (information retrieval); Computer science; Linguistics; Natural language processing; Artificial intelligence; Materials science; Philosophy","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00252166,0.002340433,0.001254423,0.00199211,0.0009082344,0.002498183,0.002618867,0.002311331,0.007764936],"category_scores_gemma":[0.007568032,0.0003959689,0.001254004,0.001045367,0.001086853,0.003812132,0.003271962,0.003406552,0.002969238],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001378797,"about_ca_system_score_gemma":0.001535734,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005875855,"about_ca_topic_score_gemma":0.01298054,"domain_scores_codex":[0.9985977,0.000291766,0.00006033106,0.0005365764,0.0003077519,0.0002058592],"domain_scores_gemma":[0.997241,0.001296291,0.0001440193,0.0004610395,0.0006837735,0.0001739174],"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":[0.001427287,0.0008789004,0.004605728,0.0005397721,0.000233563,0.0005049806,0.000323168,0.08681735,0.05203919,0.01310361,0.03053067,0.8089957],"study_design_scores_gemma":[0.00005139404,0.0002116131,0.00123699,0.00004281272,0.00006573027,0.0001571981,0.0001891692,0.9490126,0.02618345,0.01900877,0.003806857,0.00003349218],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.180974,0.001316831,0.7809234,0.001271136,0.0003549126,0.0004066983,0.001865933,0.0153627,0.01752441],"genre_scores_gemma":[0.6826826,0.0003398065,0.2947278,0.00133191,0.0002174526,0.0002986536,0.006505821,0.001303144,0.01259284],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.007764936,"threshold_uncertainty_score":0.0259763,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02464168471662963,"score_gpt":0.3549871254119355,"score_spread":0.3303454406953059,"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."}}