{"id":"W4389518385","doi":"10.18653/v1/2023.arabicnlp-1.1","title":"Violet: A Vision-Language Model for Arabic Image Captioning with Gemini Decoder","year":2023,"lang":"en","type":"article","venue":"","topic":"Multimodal Machine Learning Applications","field":"Computer Science","cited_by":9,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of British Columbia","funders":"Alliance de recherche numérique du Canada; Social Sciences and Humanities Research Council of Canada; Natural Sciences and Engineering Research Council of Canada; Canada Research Chairs","keywords":"Closed captioning; Computer science; Fluency; Natural language processing; Artificial intelligence; Arabic; Language model; Encoder; Image (mathematics); Generative model; Speech recognition; Generative grammar; Linguistics","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":[0.0002552684,0.0001240696,0.0001146376,0.0001348852,0.0002198968,0.0001645297,0.0005411517,0.00003915163,0.00001786771],"category_scores_gemma":[0.00005412419,0.00009802474,0.00004660842,0.000489696,0.00003085616,0.0003024614,0.0001603089,0.0001132542,0.0003173376],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002576061,"about_ca_system_score_gemma":0.00005253158,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001249636,"about_ca_topic_score_gemma":0.00003151642,"domain_scores_codex":[0.9989207,0.00001779964,0.0001518211,0.0004388939,0.0001798454,0.0002908781],"domain_scores_gemma":[0.9990386,0.0001495902,0.00005955341,0.0005867447,0.00008194044,0.00008351441],"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":[0.00003804869,0.0002116248,0.002505833,0.00009724878,0.00006916108,0.0000205465,0.01211245,0.7216381,0.04029589,0.1100953,0.008789442,0.1041264],"study_design_scores_gemma":[0.0003147416,0.00003594713,0.004873402,0.00001461249,0.000004307827,0.000003524679,0.00006010356,0.9926086,0.0002407073,0.001592208,0.00009882624,0.0001530311],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1008736,0.000007397939,0.8940805,0.002261915,0.00002187818,0.0003153806,0.00000448503,0.0008013449,0.001633608],"genre_scores_gemma":[0.5824144,0.000001007625,0.4153442,0.0002301029,0.00002112413,0.0002013403,0.00001211088,0.00001547601,0.001760298],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.4815408,"threshold_uncertainty_score":0.4078839,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01223930283923168,"score_gpt":0.3186121580308731,"score_spread":0.3063728551916414,"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."}}