{"id":"W4367727879","doi":"10.1039/d2dd00146b","title":"Calibration and generalizability of probabilistic models on low-data chemical datasets with DIONYSUS","year":2023,"lang":"en","type":"article","venue":"Digital Discovery","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":48,"is_retracted":false,"has_abstract":true,"ca_institutions":"Canadian Institute for Advanced Research; Vector Institute; University of Toronto","funders":"Natural Sciences and Engineering Research Council of Canada; Alliance de recherche numérique du Canada; University of Toronto; Canada Foundation for Innovation; Government of Ontario; Government of Canada; Vector Institute; Canadian Institute for Advanced Research","keywords":"Generalizability theory; Probabilistic logic; Calibration; Computer science; Statistical model; Artificial intelligence; Data mining; Machine learning; Statistics; Mathematics","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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002611035,0.0001492106,0.0001873557,0.00007981699,0.00003611435,0.0004353963,0.0006972476,0.00003431626,6.494062e-7],"category_scores_gemma":[0.0002189957,0.0001188685,0.00002825722,0.0004846402,0.0001403375,0.005003999,0.0009739293,0.00006998249,0.000003424938],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002902962,"about_ca_system_score_gemma":0.000132253,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0000118449,"about_ca_topic_score_gemma":0.000001767965,"domain_scores_codex":[0.9984402,0.00006349065,0.0002513222,0.0006343633,0.0004251465,0.000185462],"domain_scores_gemma":[0.9983309,0.0004474327,0.00008158312,0.001031516,0.00003552119,0.00007305935],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002407343,0.0006737456,0.001077287,0.0003785435,0.00008418013,0.00004687826,0.0004165948,0.6135053,0.001009712,0.3537398,0.003481271,0.0253459],"study_design_scores_gemma":[0.0002474661,0.00007011937,0.001573774,0.00004369074,0.000005284445,0.00000701425,0.00001078176,0.9312072,0.0008429894,0.06580029,0.0000265843,0.0001648035],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5972073,0.00001388417,0.400039,0.0002103963,0.00005859517,0.0002002866,0.001913981,0.0000883356,0.0002681596],"genre_scores_gemma":[0.9937193,0.000003690617,0.003556878,0.0000686612,0.00002596671,0.000011482,0.002578932,0.00001115422,0.00002392142],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.396512,"threshold_uncertainty_score":0.4847315,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05087454827613958,"score_gpt":0.2897582750390798,"score_spread":0.2388837267629402,"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."}}