{"id":"W4383560362","doi":"10.54254/2755-2721/5/20230511","title":"To describe the content of image: The view from image captioning","year":2023,"lang":"en","type":"article","venue":"Applied and Computational Engineering","topic":"Multimodal Machine Learning Applications","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Closed captioning; Computer science; Natural language; Field (mathematics); Artificial intelligence; Natural language processing; Task (project management); Image (mathematics); Domain (mathematical analysis); Perspective (graphical); 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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0025801,0.001561343,0.0008020286,0.002891315,0.0006989947,0.004956685,0.002023357,0.002552517,0.00727852],"category_scores_gemma":[0.01337814,0.0006143539,0.0008508352,0.002168659,0.003800564,0.009760413,0.003241947,0.004668877,0.004538908],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001483251,"about_ca_system_score_gemma":0.0008972445,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00194793,"about_ca_topic_score_gemma":0.001899328,"domain_scores_codex":[0.9982384,0.0007780187,0.0001033726,0.0003512063,0.0004521692,0.0000767434],"domain_scores_gemma":[0.9936207,0.003108701,0.0003101153,0.001330308,0.001405861,0.0002243117],"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":[0.000236556,0.00006516411,0.0007651845,0.00288063,0.00008398047,0.000247579,0.001661236,0.008724152,0.01645166,0.1180326,0.08857596,0.7622753],"study_design_scores_gemma":[0.00004418105,0.0002801841,0.002140255,0.001574837,0.0001147314,0.002859909,0.001809532,0.08043665,0.04897876,0.2353226,0.6262411,0.0001972084],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.007582892,0.05976912,0.8747844,0.01792002,0.003096212,0.0003965575,0.001587949,0.003075166,0.03178772],"genre_scores_gemma":[0.1016664,0.0503032,0.8073982,0.006596091,0.004210678,0.0005382861,0.004374519,0.001863635,0.02304904],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.00727852,"threshold_uncertainty_score":0.02434909,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01803443695616382,"score_gpt":0.234893111473301,"score_spread":0.2168586745171372,"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."}}