{"id":"W4388509054","doi":"10.1101/2023.11.08.566195","title":"ChatGPT usage in the Reactome curation process","year":2023,"lang":"en","type":"preprint","venue":"bioRxiv (Cold Spring Harbor Laboratory)","topic":"Artificial Intelligence in Healthcare and Education","field":"Medicine","cited_by":13,"is_retracted":false,"has_abstract":true,"ca_institutions":"Ontario Institute for Cancer Research","funders":"National Institutes of Health; University of Toronto; European Bioinformatics Institute","keywords":"Computer science; Data curation; Annotation; Workflow; Process (computing); Context (archaeology); Pipeline (software); Automation; Data science; Information retrieval; Software engineering; Artificial intelligence; Database; Programming language; Engineering","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":[],"category_scores_codex":[0.03286838,0.003130342,0.001885915,0.003628616,0.002208555,0.005064105,0.003985042,0.00267749,0.01368477],"category_scores_gemma":[0.08052395,0.00175593,0.003565203,0.00284686,0.001839009,0.00501579,0.008389656,0.004999432,0.01232705],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002892382,"about_ca_system_score_gemma":0.004686489,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004479853,"about_ca_topic_score_gemma":0.005650339,"domain_scores_codex":[0.9821953,0.00915931,0.001570964,0.003276692,0.003128928,0.0006688781],"domain_scores_gemma":[0.9108134,0.05629931,0.002474175,0.01862762,0.009713395,0.002072045],"domain_codex":null,"domain_gemma":"methods","domain_candidate":"methods","domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.006814964,0.000732467,0.01577619,0.01373541,0.00196699,0.005998979,0.03389685,0.03581925,0.1718899,0.05062271,0.2516945,0.4110518],"study_design_scores_gemma":[0.0005944357,0.0006846109,0.006729046,0.001465006,0.0006270593,0.002127007,0.002378282,0.1867223,0.1547018,0.06054525,0.5824946,0.0009306048],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01858215,0.0005663345,0.6392923,0.002766891,0.0009277598,0.001045769,0.01425878,0.3125383,0.01002165],"genre_scores_gemma":[0.162986,0.0007598668,0.7066577,0.003276126,0.0003531681,0.003690488,0.04321026,0.06609398,0.01297245],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9671316,"threshold_uncertainty_score":0.1738266,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1201767186247262,"score_gpt":0.3720917465537547,"score_spread":0.2519150279290285,"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."}}