{"id":"W4367672600","doi":"10.36227/techrxiv.22133894.v2","title":"Image Captioning for the Visually Impaired and Blind: A Recipe for Low-Resource Languages","year":2023,"lang":"en","type":"preprint","venue":"","topic":"Multimodal Machine Learning Applications","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"","keywords":"Closed captioning; Computer science; Visually impaired; Assistive technology; Hearing impaired; Human–computer interaction; Artificial intelligence; Speech recognition; Computer vision; Image (mathematics); Natural language processing; Audiology","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":[0.0008977535,0.0002285988,0.0002300229,0.0001210771,0.0004157365,0.0006227339,0.001211244,0.0001514955,0.000005587868],"category_scores_gemma":[0.0005085359,0.0001700099,0.0001534479,0.0001498844,0.00006069514,0.00008128434,0.001339553,0.0003526745,0.00001903233],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003763652,"about_ca_system_score_gemma":0.0000938583,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0008527261,"about_ca_topic_score_gemma":0.0001190971,"domain_scores_codex":[0.9984515,0.00005716511,0.0002801378,0.0007368877,0.000172956,0.0003013607],"domain_scores_gemma":[0.9966539,0.001978046,0.0002005728,0.0009502156,0.0001411131,0.00007622308],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0003856096,0.0005411756,0.001842382,0.003546009,0.001274336,0.0000136325,0.02899155,0.07907676,0.01397028,0.2200149,0.06599492,0.5843484],"study_design_scores_gemma":[0.0005102388,0.00004318146,0.004482293,0.00006413658,0.00003270584,0.000003457613,0.0001579542,0.9865685,0.000132986,0.005308099,0.002446427,0.0002500887],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.008997686,0.0001254279,0.9761404,0.01131838,0.0001375822,0.00221762,0.00005021579,0.000781261,0.0002313819],"genre_scores_gemma":[0.2404862,0.00002329331,0.7504056,0.0006186867,0.0004728082,0.004632546,0.0001405003,0.00008626191,0.003134159],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.9074917,"threshold_uncertainty_score":0.6932803,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03894768839832212,"score_gpt":0.3599525522228245,"score_spread":0.3210048638245024,"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."}}