{"id":"W3197433610","doi":"10.1038/s41592-021-01256-7","title":"Reproducibility standards for machine learning in the life sciences","year":2021,"lang":"en","type":"article","venue":"Nature Methods","topic":"Cell Image Analysis Techniques","field":"Biochemistry, Genetics and Molecular Biology","cited_by":228,"is_retracted":false,"has_abstract":false,"ca_institutions":"Vector Institute; Princess Margaret Cancer Centre; University of Toronto; University Health Network","funders":"University of Colorado School of Medicine, Anschutz Medical Campus; National Institute of General Medical Sciences; Perelman School of Medicine, University of Pennsylvania; National Cancer Institute; University of Toronto; U.S. Department of Health and Human Services; National Institutes of Health; Cancer Research UK; Government of Canada; University of Washington; National Human Genome Research Institute; Johns Hopkins Bloomberg School of Public Health; Cancer Research UK Cambridge Institute, University of Cambridge; University of Pennsylvania; Natural Sciences and Engineering Research Council of Canada; Johns Hopkins University","keywords":"Workflow; Computer science; Automation; Reproducibility; Data science; Artificial intelligence; Software engineering; Code (set theory); Machine learning; Programming language; Database; Engineering; Chemistry","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":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.01538814,0.00008285858,0.000135952,0.00003383322,0.0001098467,0.00003052103,0.0002414342,0.0001793114,0.00001474292],"category_scores_gemma":[0.02253645,0.00005709505,0.0001141228,0.0003346179,0.00008430691,0.000002822822,0.00009812398,0.0003541643,1.01303e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001111733,"about_ca_system_score_gemma":0.0002220619,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001077908,"about_ca_topic_score_gemma":0.00007129842,"domain_scores_codex":[0.997881,0.0008997966,0.0001538314,0.0007500073,0.0001719914,0.0001433935],"domain_scores_gemma":[0.9987447,0.0001583924,0.00005304445,0.0008073258,0.0002159556,0.00002055331],"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":[0.0000533407,0.00007656228,0.00627779,0.00003393754,0.00004334224,0.000004553364,0.0001064726,0.00005054027,0.9229004,0.0002348163,0.009172603,0.06104568],"study_design_scores_gemma":[0.00009801448,0.00007247852,0.0007074631,0.000003634044,0.00001926419,0.000005168032,0.00007313825,0.000232073,0.6745641,0.0007736974,0.3233812,0.00006969136],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.1971688,0.1356008,0.6377586,0.008485104,0.000246453,0.0009828295,0.00004276292,0.0001039773,0.01961059],"genre_scores_gemma":[0.486377,0.0004645892,0.5097776,0.002341904,0.0002328141,0.00004207985,0.00008320059,0.00001286849,0.0006679442],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.3142086,"threshold_uncertainty_score":0.9856972,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02233273429208008,"score_gpt":0.4364528823126647,"score_spread":0.4141201480205846,"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."}}