{"id":"W3036441178","doi":"10.1016/j.chempr.2020.05.014","title":"Learning to Make Chemical Predictions: The Interplay of Feature Representation, Data, and Machine Learning Methods","year":2020,"lang":"en","type":"article","venue":"Chem","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":118,"is_retracted":false,"has_abstract":false,"ca_institutions":"Queen's University","funders":"National Institute of General Medical Sciences; National Institutes of Health","keywords":"Machine learning; Artificial intelligence; Computer science; Property (philosophy); Chemical space; Feature (linguistics); Representation (politics); Perspective (graphical); Kernel (algebra); Kernel method; Feature learning; Focus (optics); Feature vector; Random forest; Support vector machine; Mathematics; Chemistry; Drug discovery","routes":{"ca_aff":true,"ca_fund":false,"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":[0.0007280612,0.00009503448,0.0001485202,0.00003527034,0.00009172726,0.00008929749,0.0008205195,0.00003643878,0.00000782312],"category_scores_gemma":[0.002184046,0.0000748794,0.00003105237,0.0005078205,0.00004882447,0.0002178045,0.001421915,0.0003758293,0.000003676702],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001353088,"about_ca_system_score_gemma":0.00003718144,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000009005474,"about_ca_topic_score_gemma":7.102736e-7,"domain_scores_codex":[0.9987205,0.0003537424,0.0001834651,0.0004225793,0.0002060092,0.0001137113],"domain_scores_gemma":[0.9984428,0.000929227,0.0000926174,0.0003729007,0.00007076134,0.0000916514],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00007972147,0.00006322105,0.008621107,0.0001268045,0.0001758682,0.000008185813,0.03526328,0.1134274,0.07116766,0.004679855,0.006637668,0.7597492],"study_design_scores_gemma":[0.0001511852,0.00004725746,0.001731646,0.00001826421,0.00001260788,0.00001856052,0.0003161839,0.9519055,0.02646217,0.0005364558,0.01870843,0.00009168706],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.02378245,0.0002705069,0.9646195,0.01064637,0.00008920576,0.0001239925,0.000006869575,0.00008936926,0.0003717611],"genre_scores_gemma":[0.3777433,0.00001534508,0.6212695,0.0006050441,0.0001305505,0.00001062555,0.00004322389,0.00001153863,0.0001708549],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.8384782,"threshold_uncertainty_score":0.3053493,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0505661784080373,"score_gpt":0.3798396244581497,"score_spread":0.3292734460501124,"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."}}