{"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000810004,0.001197878,0.0006207488,0.001004789,0.0003760653,0.0009539957,0.00193988,0.0009773346,0.01028438],"category_scores_gemma":[0.001862014,0.0003689901,0.0007856076,0.0004962935,0.0003851145,0.001477251,0.00145597,0.001198061,0.008326942],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006978602,"about_ca_system_score_gemma":0.0008477043,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005598105,"about_ca_topic_score_gemma":0.009087399,"domain_scores_codex":[0.9996768,0.00005621289,0.00001964933,0.0001135155,0.00009051251,0.0000432347],"domain_scores_gemma":[0.9994215,0.0001394311,0.00003885066,0.0001053822,0.0002506408,0.00004424074],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0004264764,0.000180097,0.00146129,0.0003538746,0.0001056649,0.0001496532,0.0001175478,0.03008658,0.04298603,0.00190508,0.02540412,0.8968235],"study_design_scores_gemma":[0.00004098898,0.0002785561,0.001454621,0.00007908501,0.00006421269,0.0002483067,0.00007327543,0.9094682,0.05988917,0.003390676,0.02495876,0.00005416323],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0272543,0.001405203,0.9036877,0.0003793739,0.0002873061,0.0003484243,0.002438828,0.05881082,0.005388063],"genre_scores_gemma":[0.3596101,0.001045156,0.5903728,0.001445509,0.0002194624,0.0007456938,0.008913282,0.002018468,0.0356296],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01028438,"threshold_uncertainty_score":0.03440475,"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."}}