{"id":"W3173937618","doi":"10.18653/v1/2021.acl-short.36","title":"Enhancing Descriptive Image Captioning with Natural Language Inference","year":2021,"lang":"en","type":"article","venue":"","topic":"Multimodal Machine Learning Applications","field":"Computer Science","cited_by":15,"is_retracted":false,"has_abstract":true,"ca_institutions":"Queen's University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Closed captioning; Computer science; Inference; Natural language; Computational linguistics; Natural language processing; Natural (archaeology); Joint (building); Artificial intelligence; Linguistics; Image (mathematics); History; Engineering; Philosophy","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":[],"consensus_categories":[],"category_scores_codex":[0.00009084227,0.00009379398,0.00008828912,0.00004115036,0.0001428894,0.00022688,0.0003180973,0.00002129809,0.00006670271],"category_scores_gemma":[0.0001315288,0.00007673637,0.0000228477,0.0003893491,0.00002863082,0.0004310774,0.0001832303,0.0002285777,0.0001261831],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004296396,"about_ca_system_score_gemma":0.0001109424,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0007373218,"about_ca_topic_score_gemma":0.0004519454,"domain_scores_codex":[0.9991717,0.00005037574,0.0001045751,0.0003255517,0.0001619839,0.0001858446],"domain_scores_gemma":[0.9992108,0.0001134096,0.00004562462,0.0004150979,0.0001594473,0.00005556992],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000005881576,0.0001046096,0.007064388,0.00002918603,0.00004773547,0.0001860425,0.017193,0.0009653778,0.7656012,0.1670303,0.00009436208,0.04167794],"study_design_scores_gemma":[0.000811482,0.00008001864,0.1278304,0.0001287143,0.00001770305,0.0002245496,0.003269544,0.4991009,0.3663982,0.001011535,0.0003399432,0.0007870768],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2165108,0.00007712322,0.7762111,0.0006532438,0.00004405145,0.00005401775,4.227968e-7,0.0002548746,0.006194358],"genre_scores_gemma":[0.6300474,7.448974e-7,0.3691204,0.0001562241,0.00001715376,0.00001191481,0.000003463839,0.000004446709,0.0006381838],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.4981355,"threshold_uncertainty_score":0.3129219,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.007255126184124819,"score_gpt":0.2684051231919964,"score_spread":0.2611499970078716,"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."}}