{"id":"W3117879109","doi":"10.1039/d0qo01636e","title":"Data augmentation and transfer learning strategies for reaction prediction in low chemical data regimes","year":2021,"lang":"en","type":"article","venue":"Organic Chemistry Frontiers","topic":"Machine Learning in Materials Science","field":"Materials Science","cited_by":71,"is_retracted":false,"has_abstract":true,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"National Natural Science Foundation of China","keywords":"Chemistry; Drug discovery; Transfer of learning; Biochemical engineering; Data science; Artificial intelligence; Computer science; Biochemistry","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":[0.0007486348,0.0001380056,0.0001851718,0.00001565119,0.0001774262,0.000426659,0.0005464991,0.000107811,0.0004431828],"category_scores_gemma":[0.0004819611,0.0001500033,0.00001163819,0.0001436218,0.0001073521,0.001012152,0.0002626266,0.0001642128,0.000005432918],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00008080163,"about_ca_system_score_gemma":0.0002225572,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003082966,"about_ca_topic_score_gemma":0.00001195382,"domain_scores_codex":[0.9983902,0.00006097995,0.0002787533,0.0008179108,0.0002127114,0.0002394465],"domain_scores_gemma":[0.9990375,0.00005249143,0.00007161315,0.0007314174,0.00004613938,0.00006085288],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.00004478159,0.00002618553,0.0004033708,0.0001797453,0.000004857844,0.000003311393,0.0001554434,0.0001365241,0.9966422,0.000005659985,0.001989102,0.0004087567],"study_design_scores_gemma":[0.0008244358,0.00001505569,0.0004119434,0.0000823376,0.00003883921,0.00004545454,0.002524154,0.04577027,0.94573,0.0002066697,0.004131182,0.0002196937],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.956638,0.0001714804,0.04160159,0.0003285634,0.0004291029,0.0001400104,0.0002585561,0.0000995731,0.0003330673],"genre_scores_gemma":[0.9770935,0.00009656068,0.01713592,0.00003640881,0.0002730185,0.00001860815,0.003943253,0.0000358821,0.001366808],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.05091229,"threshold_uncertainty_score":0.6116955,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01940198101854031,"score_gpt":0.2656183214908371,"score_spread":0.2462163404722968,"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."}}