{"id":"W3116946603","doi":"10.2196/22795","title":"Adapting Bidirectional Encoder Representations from Transformers (BERT) to Assess Clinical Semantic Textual Similarity: Algorithm Development and Validation Study","year":2020,"lang":"en","type":"article","venue":"JMIR Medical Informatics","topic":"Topic Modeling","field":"Computer Science","cited_by":24,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Deutschen Konsortium für Translationale Krebsforschung; Deutsches Krebsforschungszentrum","keywords":"Computer science; Encoder; Pearson product-moment correlation coefficient; Leverage (statistics); Artificial intelligence; Semantic similarity; Natural language processing; Relationship extraction; Sentence; Test set; Ground truth; Machine learning; Information extraction; Statistics","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":[0.003812643,0.001597128,0.0009923778,0.001757876,0.0004084932,0.001076829,0.001542934,0.00142587,0.00278423],"category_scores_gemma":[0.01382129,0.0003805803,0.0007559796,0.001013116,0.000410931,0.001325949,0.00141615,0.001690786,0.001312362],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001358186,"about_ca_system_score_gemma":0.002192266,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01381257,"about_ca_topic_score_gemma":0.01405129,"domain_scores_codex":[0.9985139,0.0006410738,0.0001047034,0.0003041671,0.0002935447,0.0001426194],"domain_scores_gemma":[0.9942144,0.003986137,0.0002151477,0.0003974481,0.001016355,0.0001706213],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0007723265,0.0005432544,0.0100668,0.0002848397,0.0002429747,0.0001676447,0.0001479795,0.2762241,0.004133924,0.002803874,0.01098503,0.6936272],"study_design_scores_gemma":[0.00004604927,0.0001245076,0.0007761682,0.00001838812,0.00002427089,0.00006141814,0.00004605959,0.994561,0.002513529,0.001201552,0.0006162555,0.00001082167],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.288555,0.00211683,0.6828468,0.000656785,0.0003315168,0.0008245109,0.002045656,0.01800992,0.004612993],"genre_scores_gemma":[0.6842617,0.0004813521,0.3045365,0.0002662965,0.0000863715,0.0005015462,0.006316203,0.0004383204,0.003111744],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01381257,"threshold_uncertainty_score":0.02746433,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1256573497326064,"score_gpt":0.3820119822168946,"score_spread":0.2563546324842882,"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."}}