{"id":"W4318465586","doi":"10.1007/s10032-023-00427-w","title":"Large-scale genealogical information extraction from handwritten Quebec parish records","year":2023,"lang":"en","type":"article","venue":"International Journal on Document Analysis and Recognition (IJDAR)","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":13,"is_retracted":false,"has_abstract":false,"ca_institutions":"Université du Québec à Chicoutimi","funders":"Association Nationale de la Recherche et de la Technologie","keywords":"Computer science; Workflow; Consistency (knowledge bases); Scale (ratio); Sample (material); Information extraction; Population; Artificial intelligence; Natural language processing; Information retrieval; Data mining; Database; Geography; Medicine; Cartography","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.0006666536,0.000154919,0.0002079509,0.001020606,0.0002249963,0.001602959,0.0004888465,0.0001175167,0.0003692618],"category_scores_gemma":[0.00008879191,0.0001228037,0.0002076481,0.0007564221,0.00002251183,0.00242849,0.0001505126,0.0003456644,0.0001664868],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001571578,"about_ca_system_score_gemma":0.00003906472,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0004981934,"about_ca_topic_score_gemma":0.0005345686,"domain_scores_codex":[0.9981197,0.0001054578,0.0005220373,0.0002596851,0.0007901908,0.0002029184],"domain_scores_gemma":[0.9987554,0.0001113996,0.0004081313,0.0001469387,0.0004547853,0.0001233756],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0001183902,0.0001477795,0.004409354,0.000005822922,0.001249194,0.0001184789,0.001608756,0.00009945362,0.0005317415,0.001204816,0.008225637,0.9822806],"study_design_scores_gemma":[0.005453192,0.0007361511,0.1387794,0.0006550938,0.001496473,0.0003890081,0.001984913,0.1622871,0.02707503,0.5588711,0.09988011,0.002392362],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2954084,0.0001988706,0.6965792,0.005582562,0.001196784,0.0001184966,0.00006573413,0.000404079,0.0004459352],"genre_scores_gemma":[0.9294596,0.001643642,0.06415938,0.002070535,0.0007624999,0.00002833557,0.001073634,0.00001132139,0.0007909949],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9798882,"threshold_uncertainty_score":0.9994335,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01208182651710639,"score_gpt":0.2916883980205446,"score_spread":0.2796065715034383,"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."}}