{"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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001934328,0.00008314702,0.00008806521,0.00004750644,0.0001446883,0.00009573499,0.0003629025,0.00001447051,0.000003120319],"category_scores_gemma":[0.00003127954,0.00005664667,0.00002340202,0.000401488,0.00002802389,0.00006302103,0.000197325,0.0001126252,0.00005419622],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001074936,"about_ca_system_score_gemma":0.00001495961,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001498844,"about_ca_topic_score_gemma":9.546879e-7,"domain_scores_codex":[0.9993644,0.00001213772,0.000154626,0.0001775586,0.0001711386,0.0001200836],"domain_scores_gemma":[0.9992006,0.0004709657,0.00003978189,0.0001975087,0.0000512561,0.00003982079],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000001570932,0.000007522605,0.00005975877,0.0000109443,0.00002279116,6.116461e-7,0.001279118,0.7314239,0.02422987,0.2249698,0.0001283703,0.01786574],"study_design_scores_gemma":[0.0001037402,0.000005146471,0.1324435,0.00001251833,0.000004627498,0.000001941734,0.00006567689,0.8623604,0.0002553093,0.004333514,0.0003393533,0.00007424738],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1504046,0.00004088434,0.8455959,0.003417842,0.00004237112,0.0001876924,0.000004859578,0.0001682048,0.0001376586],"genre_scores_gemma":[0.9013194,0.000003173994,0.09831923,0.000207866,0.00003721417,0.00008596743,0.00001106947,0.000007757054,0.00000829526],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.7509148,"threshold_uncertainty_score":0.2309984,"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."}}