{"id":"W4402671279","doi":"10.18653/v1/2024.arabicnlp-1.19","title":"Qalam: A Multimodal LLM for Arabic Optical Character and Handwriting Recognition","year":2024,"lang":"en","type":"article","venue":"","topic":"Handwritten Text Recognition Techniques","field":"Computer Science","cited_by":9,"is_retracted":false,"has_abstract":true,"ca_institutions":"","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":"Computer science; Character (mathematics); Arabic; Handwriting; Optical character recognition; Intelligent character recognition; Character recognition; Handwriting recognition; Speech recognition; Natural language processing; Artificial intelligence; Feature extraction; Linguistics; Image (mathematics); Mathematics","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000399,0.0001409492,0.0001463612,0.0001736288,0.0001004301,0.0006425207,0.0001806735,0.00009964483,0.00005858196],"category_scores_gemma":[0.00009254256,0.000122762,0.00007795196,0.0001779969,0.00004562361,0.0008904043,0.00010806,0.0001434325,0.0001151357],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002587308,"about_ca_system_score_gemma":0.00003393725,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000005984029,"about_ca_topic_score_gemma":0.00000351284,"domain_scores_codex":[0.9988421,0.00002619002,0.0002354797,0.0004860769,0.0001375641,0.0002725815],"domain_scores_gemma":[0.9992823,0.0003167946,0.00002372057,0.0001638969,0.0001112909,0.0001020094],"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":[0.000005347589,0.00004009143,0.00001732745,0.000097376,0.00001470625,0.00001418334,0.0001485539,2.85899e-8,0.006259234,0.01499467,0.0007240635,0.9776844],"study_design_scores_gemma":[0.001394244,0.000883762,0.001410214,0.001011215,0.00006593023,0.0006070049,0.0001055042,0.4792126,0.3595277,0.135258,0.01926375,0.001259998],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03320898,0.0001244442,0.9589155,0.00197416,0.0002098596,0.0004566782,0.00001320163,0.001344273,0.003752874],"genre_scores_gemma":[0.578208,0.00004808673,0.4200276,0.0006057155,0.0002336574,0.0002772555,0.00002391329,0.00002011726,0.000555596],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9764244,"threshold_uncertainty_score":0.6195842,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02340246789728012,"score_gpt":0.2742892707726949,"score_spread":0.2508868028754148,"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."}}