{"id":"W3119848647","doi":"10.2196/23101","title":"Incorporating Domain Knowledge Into Language Models by Using Graph Convolutional Networks for Assessing Semantic Textual Similarity: Model Development and Performance Comparison","year":2021,"lang":"en","type":"article","venue":"JMIR Medical Informatics","topic":"Machine Learning in Healthcare","field":"Computer Science","cited_by":17,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"U.S. National Library of Medicine; National Center for Advancing Translational Sciences; National Institute of Mental Health; National Institutes of Health; U.S. Department of Veterans Affairs","keywords":"Computer science; Natural language processing; Artificial intelligence; Semantic similarity; Information retrieval; Encoder; Automatic summarization; Sentence; Documentation; Unified Medical Language System; Graph; Knowledge base; Theoretical computer science","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":[0.003663942,0.002132332,0.0007810352,0.001811433,0.0004756011,0.001126377,0.002092117,0.001446239,0.001678817],"category_scores_gemma":[0.008597646,0.0005925095,0.001464035,0.001170101,0.0004616789,0.002015498,0.001058581,0.002866906,0.0006901819],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.003063236,"about_ca_system_score_gemma":0.002304746,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0504459,"about_ca_topic_score_gemma":0.05360837,"domain_scores_codex":[0.9990025,0.0004127282,0.00006559169,0.0002805632,0.0001289436,0.0001095602],"domain_scores_gemma":[0.9955832,0.002965321,0.0001745018,0.0003154478,0.0008396664,0.0001218958],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0003725508,0.0003326987,0.004268267,0.0001348832,0.0002739029,0.00007197447,0.00006075302,0.8614563,0.001766724,0.00100102,0.002511781,0.1277492],"study_design_scores_gemma":[0.000008806225,0.0000305966,0.0001668563,0.000006848818,0.00002191172,0.000005775626,0.000006462198,0.998524,0.0005649582,0.000533563,0.0001256768,0.000004561443],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4200065,0.003471462,0.5548302,0.002078201,0.0003816276,0.000739674,0.002728274,0.01014105,0.005623053],"genre_scores_gemma":[0.814831,0.0009092006,0.1754092,0.0004079344,0.00008932717,0.000551931,0.004498831,0.0002730941,0.003029548],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.0504459,"threshold_uncertainty_score":0.1003045,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03909626719284735,"score_gpt":0.3488209884979172,"score_spread":0.3097247213050698,"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."}}