{"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001004007,0.002381668,0.0008331291,0.001040851,0.0004944724,0.00246296,0.001782622,0.001958462,0.02277143],"category_scores_gemma":[0.01010028,0.0007049573,0.001422551,0.0007856936,0.0004799876,0.005994651,0.002415179,0.002796496,0.009521676],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001056404,"about_ca_system_score_gemma":0.001335835,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005711624,"about_ca_topic_score_gemma":0.007082063,"domain_scores_codex":[0.9988331,0.0004158947,0.00005779781,0.0004027,0.0002077411,0.00008280158],"domain_scores_gemma":[0.9957918,0.002583749,0.0002211984,0.0006465783,0.0006062603,0.0001505054],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.003393983,0.001504663,0.004519568,0.001348769,0.0003645976,0.0008285132,0.0006407966,0.1747491,0.06132601,0.03104734,0.07759105,0.6426857],"study_design_scores_gemma":[0.000128991,0.000173821,0.000608492,0.00007446582,0.0001322662,0.000138813,0.0001976837,0.9482526,0.01627772,0.02737219,0.00658905,0.00005392015],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.1305747,0.001935196,0.7884946,0.002382381,0.001707445,0.0004907749,0.00866506,0.03794383,0.02780605],"genre_scores_gemma":[0.81229,0.00101301,0.1519224,0.001312798,0.00036217,0.0002852508,0.01160549,0.002541148,0.01866781],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.02277143,"threshold_uncertainty_score":0.07617807,"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."}}