{"id":"W4399668246","doi":"10.1145/3661167.3661172","title":"The Promise and Challenges of Using LLMs to Accelerate the Screening Process of Systematic Reviews","year":2024,"lang":"en","type":"article","venue":"","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":22,"is_retracted":false,"has_abstract":true,"ca_institutions":"Dalhousie University","funders":"Strategic Research Council; Killam Trusts","keywords":"Process (computing); Systematic review; Computer science; Risk analysis (engineering); Medicine; MEDLINE; Political science","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":["metaresearch"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.4747727,0.003327138,0.00472166,0.01182363,0.002406956,0.01273088,0.004738777,0.004289876,0.007691504],"category_scores_gemma":[0.8019078,0.003831252,0.006587993,0.01067367,0.003470387,0.01873494,0.009267641,0.006462672,0.003521335],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.009207926,"about_ca_system_score_gemma":0.03090198,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005860028,"about_ca_topic_score_gemma":0.01588474,"domain_scores_codex":[0.4662976,0.4791699,0.02605627,0.009178506,0.0179633,0.001334415],"domain_scores_gemma":[0.03051294,0.8927956,0.02338712,0.03188363,0.01954773,0.001873008],"domain_codex":"methods","domain_gemma":"methods","domain_candidate":"methods","domain_consensus":"methods","study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.00630826,0.0008824566,0.0126411,0.07601871,0.004505684,0.0002601479,0.01529404,0.01162647,0.01058858,0.009831676,0.02494404,0.827099],"study_design_scores_gemma":[0.01431626,0.01564568,0.0876583,0.1077236,0.01556573,0.002930269,0.01305178,0.2360922,0.04296932,0.2040825,0.256183,0.00378127],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1271392,0.07388745,0.5941712,0.1400692,0.00350829,0.01809102,0.004067481,0.02698063,0.01208558],"genre_scores_gemma":[0.07990169,0.005340365,0.904202,0.003092641,0.0005565027,0.005263962,0.0006763358,0.0005581839,0.0004083085],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.5252273,"threshold_uncertainty_score":0.6476988,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1590595772908095,"score_gpt":0.3678787959978982,"score_spread":0.2088192187070887,"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."}}