{"id":"W3096461368","doi":"10.2196/18953","title":"The Impact of Pretrained Language Models on Negation and Speculation Detection in Cross-Lingual Medical Text: Comparative Study","year":2020,"lang":"en","type":"article","venue":"JMIR Medical Informatics","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":30,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Ministerio de Economía y Competitividad","keywords":"Computer science; Negation; Natural language processing; Artificial intelligence; Speculation; Conditional random field; Programming language","routes":{"ca_aff":false,"ca_fund":false,"ca_venue":true,"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":[0.008432997,0.002385012,0.001263587,0.001912343,0.0008240666,0.002981859,0.001995193,0.001832522,0.003768162],"category_scores_gemma":[0.02796379,0.0007134499,0.00181093,0.001391463,0.0008693381,0.005645773,0.002156473,0.002794592,0.002896902],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001786432,"about_ca_system_score_gemma":0.001375917,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.02020353,"about_ca_topic_score_gemma":0.01682046,"domain_scores_codex":[0.9960598,0.001951315,0.0003510295,0.001029179,0.0003489451,0.0002597102],"domain_scores_gemma":[0.967477,0.02732892,0.0006905801,0.001861356,0.002226468,0.0004156533],"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":[0.006310999,0.003365293,0.05185769,0.003152027,0.003815731,0.00204438,0.001853326,0.1371415,0.00962503,0.00219558,0.02813473,0.7505038],"study_design_scores_gemma":[0.0002214058,0.001087901,0.02101726,0.0002884192,0.001241021,0.0006634808,0.00136286,0.9583394,0.007577683,0.003583259,0.00442481,0.0001925104],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9324516,0.01357336,0.03170064,0.00198788,0.001111498,0.0002546533,0.004509058,0.004320554,0.01009083],"genre_scores_gemma":[0.9667009,0.002224636,0.01712146,0.0003475093,0.0002068452,0.0001586551,0.01079191,0.0003537933,0.002094233],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.02020353,"threshold_uncertainty_score":0.04459852,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03040752701534484,"score_gpt":0.3791577345735764,"score_spread":0.3487502075582316,"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."}}