{"id":"W4401042805","doi":"10.18653/v1/2024.findings-naacl.267","title":"Semantically-Prompted Language Models Improve Visual Descriptions","year":2024,"lang":"en","type":"article","venue":"","topic":"Multimodal Machine Learning Applications","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":false,"ca_institutions":"","funders":"Natural Sciences and Engineering Research Council of Canada; Alberta Machine Intelligence Institute","keywords":"Computer science; Natural language processing; Artificial intelligence; Visual language; Programming language; Linguistics","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0001611707,0.0001159662,0.00009332584,0.0001237671,0.0001034899,0.0004080505,0.0005269359,0.00005164895,0.00009522383],"category_scores_gemma":[0.00002719611,0.00009737258,0.00006507676,0.0004632753,0.00002977052,0.0004959074,0.0002428905,0.0002327339,0.0009174235],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003486961,"about_ca_system_score_gemma":0.00005885699,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0004656628,"about_ca_topic_score_gemma":0.00001477913,"domain_scores_codex":[0.9989305,0.000035152,0.0001661483,0.0004432552,0.0001897722,0.000235206],"domain_scores_gemma":[0.9992826,0.00009374874,0.000015794,0.0004644601,0.00004169409,0.0001016996],"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":[7.11932e-7,0.00006587781,0.0000299839,0.00002366845,0.00002087581,0.00001558667,0.000975657,0.001029336,0.03126846,0.9056326,0.000624356,0.0603129],"study_design_scores_gemma":[0.0000590534,0.0000249834,0.0004517657,0.00001237086,0.000006345401,0.00001092839,0.00002681466,0.9859793,0.0009998468,0.01173671,0.000562744,0.0001291505],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03779262,0.0001274446,0.9431986,0.003236332,0.0001994165,0.0001833986,0.000002874811,0.001752903,0.01350639],"genre_scores_gemma":[0.8765419,0.00000261306,0.1211834,0.0002384119,0.00006649193,0.00005576126,0.0000058244,0.00001370965,0.001891862],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9849499,"threshold_uncertainty_score":0.9998605,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01238632362726287,"score_gpt":0.298642225773294,"score_spread":0.2862559021460311,"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."}}