{"id":"W3200439183","doi":"10.1609/aaai.v36i10.21322","title":"BROS: A Pre-trained Language Model Focusing on Text and Layout for Better Key Information Extraction from Documents","year":2022,"lang":"en","type":"article","venue":"Proceedings of the AAAI Conference on Artificial Intelligence","topic":"Topic Modeling","field":"Computer Science","cited_by":120,"is_retracted":false,"has_abstract":true,"ca_institutions":"Kootenay Association for Science & Technology","funders":"","keywords":"Computer science; Key (lock); Natural language processing; Task (project management); Space (punctuation); Artificial intelligence; Masking (illustration); Information extraction; Language model; Semantics (computer science); Information retrieval","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00139938,0.002441808,0.001295356,0.002517364,0.0004849295,0.001391006,0.00301589,0.00178461,0.007020789],"category_scores_gemma":[0.002824639,0.0007798419,0.001708048,0.001367958,0.000536326,0.003953326,0.001607666,0.002628723,0.007925421],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008689368,"about_ca_system_score_gemma":0.002208606,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01229339,"about_ca_topic_score_gemma":0.02268618,"domain_scores_codex":[0.9992369,0.00013984,0.00005484851,0.0003145583,0.0001537018,0.0001002712],"domain_scores_gemma":[0.9984743,0.0005983987,0.0000996888,0.0002920821,0.0004366357,0.00009887557],"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":[0.0007048042,0.0004709655,0.001820645,0.0007721414,0.0003083802,0.000304377,0.0001847924,0.05006388,0.05183018,0.00383789,0.06878596,0.8209161],"study_design_scores_gemma":[0.0001273449,0.0002765126,0.001089833,0.00005726779,0.00009322484,0.0002668659,0.0001301856,0.9455866,0.03200076,0.002793463,0.0174933,0.0000846174],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0431048,0.0021888,0.8723886,0.0006891986,0.0007860247,0.0005605002,0.005868124,0.06819866,0.006215296],"genre_scores_gemma":[0.2231523,0.001048126,0.7164442,0.001095465,0.0003057036,0.0008425877,0.03155712,0.00152986,0.02402466],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01229339,"threshold_uncertainty_score":0.02444369,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0484413630078671,"score_gpt":0.3015157536877635,"score_spread":0.2530743906798963,"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."}}