{"id":"W4404781320","doi":"10.18653/v1/2024.tsar-1.5","title":"Cochrane-auto: An Aligned Dataset for the Simplification of Biomedical Abstracts","year":2024,"lang":"en","type":"article","venue":"","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":4,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Nederlandse Organisatie voor Wetenschappelijk Onderzoek; Universiteit van Amsterdam; Canadian Institute of Steel Construction","keywords":"Computer science; Information retrieval; Data science","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0003504554,0.00007117947,0.00008030018,0.00002087181,0.00004250736,0.00002134299,0.0002110807,0.0001357432,0.0000347163],"category_scores_gemma":[0.0001949037,0.00004129425,0.0000444475,0.00006133255,0.0002215173,0.000001968287,0.00003550242,0.00003068694,0.000006694569],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000002672853,"about_ca_system_score_gemma":0.00006034303,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002938069,"about_ca_topic_score_gemma":0.00001681014,"domain_scores_codex":[0.9993536,0.00001741728,0.0001848079,0.0002127784,0.000106302,0.0001250243],"domain_scores_gemma":[0.9994633,0.000117879,0.00002911496,0.0003054052,0.00002993694,0.0000543789],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00005914951,0.00008941008,0.00001043585,0.0001131269,0.00008802255,0.000001667309,0.00004951795,0.000005603132,0.689352,0.0007118213,0.1842086,0.1253106],"study_design_scores_gemma":[0.0002025,0.0004355741,0.000671919,0.00001898115,0.00003715913,0.00001080531,0.0001814152,0.001953722,0.09445996,0.000234682,0.9017007,0.00009263039],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4443245,0.02270905,0.4870282,0.02513297,0.002933197,0.001716656,0.01431942,0.000324748,0.001511246],"genre_scores_gemma":[0.9900211,0.000174468,0.002759296,0.0002957546,0.0002265004,0.00002888908,0.006253968,0.000009638898,0.000230358],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.717492,"threshold_uncertainty_score":0.1683931,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03063519998642854,"score_gpt":0.3629747871951487,"score_spread":0.3323395872087201,"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."}}