{"id":"W4391479818","doi":"10.2196/52462","title":"Automated Category and Trend Analysis of Scientific Articles on Ophthalmology Using Large Language Models: Development and Usability Study","year":2024,"lang":"en","type":"article","venue":"JMIR Formative Research","topic":"Retinal Imaging and Analysis","field":"Medicine","cited_by":16,"is_retracted":false,"has_abstract":true,"ca_institutions":"Columbia College","funders":"National Eye Institute; National Institutes of Health; Research to Prevent Blindness","keywords":"Categorization; Computer science; Set (abstract data type); Data science; Artificial intelligence; Information retrieval; Natural language processing","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":true,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00247512,0.00009380992,0.0003262411,0.001518023,0.0002135499,0.0001070997,0.00005452436,0.00003314986,0.00003698222],"category_scores_gemma":[0.00005839907,0.0000666514,0.00005485232,0.00235664,0.0003281122,0.0001626413,0.0001160853,0.0002205092,0.000004001686],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00008887068,"about_ca_system_score_gemma":0.0000940163,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001291792,"about_ca_topic_score_gemma":0.00003141845,"domain_scores_codex":[0.9983375,0.0002890733,0.0002790567,0.0003063426,0.0005089436,0.0002790638],"domain_scores_gemma":[0.999311,0.0002042725,0.00003144686,0.0002062671,0.0001430451,0.0001039672],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0006634487,0.005502328,0.3594376,0.002362528,0.008060358,0.000987621,0.5494063,0.000704609,0.02739673,0.0008736077,0.0003376367,0.04426725],"study_design_scores_gemma":[0.0003454165,0.0002995616,0.1719009,0.0000923581,0.0002960541,0.00002591959,0.02619478,0.798622,0.002055601,0.00007824272,0.00001486084,0.00007425641],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9987865,0.0004275495,0.000052616,0.00006752075,0.00001339335,0.0002171946,0.00001581674,0.00004772875,0.0003716733],"genre_scores_gemma":[0.9996237,0.000003357007,0.0001437253,0.000002908371,0.000005317735,0.00001833866,0.00002408622,0.000006833166,0.0001717381],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.7979174,"threshold_uncertainty_score":0.2717965,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1227048562417544,"score_gpt":0.4740777505475778,"score_spread":0.3513728943058234,"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."}}