{"id":"W2914453913","doi":"10.1101/526244","title":"Towards reliable named entity recognition in the biomedical domain","year":2019,"lang":"en","type":"preprint","venue":"bioRxiv (Cold Spring Harbor Laboratory)","topic":"Topic Modeling","field":"Computer Science","cited_by":9,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto","funders":"Compute Canada; National Institutes of Health; Nvidia","keywords":"Conditional random field; CRFS; Overfitting; Computer science; Artificial intelligence; Dropout (neural networks); Named-entity recognition; Transfer of learning; Machine learning; Sequence labeling; Deep learning; Natural language processing; Task (project management); Multi-task learning; Regularization (linguistics); Generalization; Artificial neural network","routes":{"ca_aff":true,"ca_fund":true,"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":[0.01281725,0.002037314,0.001421341,0.003943541,0.0008020824,0.0030892,0.002737934,0.003170164,0.004827256],"category_scores_gemma":[0.03342144,0.0006779753,0.001322071,0.003555115,0.001287941,0.009203177,0.004300789,0.003654302,0.01150519],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009224596,"about_ca_system_score_gemma":0.002308372,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003093888,"about_ca_topic_score_gemma":0.002510755,"domain_scores_codex":[0.9935941,0.002599631,0.0004637135,0.001675989,0.001413271,0.0002533191],"domain_scores_gemma":[0.9781576,0.01057768,0.001601314,0.004586093,0.004502003,0.0005753157],"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":[0.001066934,0.0003957301,0.01070845,0.002509419,0.0003388867,0.0006994046,0.0006300411,0.1060251,0.03703737,0.029799,0.1226907,0.688099],"study_design_scores_gemma":[0.0000743288,0.0002178331,0.00441912,0.0004351473,0.000143508,0.000694094,0.0005072737,0.7966716,0.06739677,0.05918139,0.07012901,0.0001299715],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.04101272,0.008789764,0.8926106,0.006469633,0.0007762248,0.0002778396,0.01077898,0.03180419,0.007480014],"genre_scores_gemma":[0.2840077,0.003914294,0.6536763,0.001899484,0.0006395397,0.0003102074,0.04866145,0.001202622,0.005688419],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01281725,"threshold_uncertainty_score":0.06778491,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02133758801382152,"score_gpt":0.2280069111405203,"score_spread":0.2066693231266988,"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."}}