{"id":"W2902154261","doi":"10.1038/s41467-018-07619-7","title":"Why rankings of biomedical image analysis competitions should be interpreted with care","year":2018,"lang":"en","type":"article","venue":"Nature Communications","topic":"Radiomics and Machine Learning in Medical Imaging","field":"Medicine","cited_by":367,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"National Institute of Neurological Disorders and Stroke; Engineering and Physical Sciences Research Council; Medical Research Council; Ministry of Science and Technology, Taiwan; National Institute of Biomedical Imaging and Bioengineering; Grantová Agentura České Republiky; Deutsches Krebsforschungszentrum; Deutsche Forschungsgemeinschaft; Australian Research Council; Wellcome Trust; Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung; National Institutes of Health; National Science Foundation","keywords":"Ranking (information retrieval); Computer science; Rank (graph theory); Data science; Quality (philosophy); Information retrieval; Image (mathematics); Data mining; Artificial intelligence; Mathematics","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":[0.0001951471,0.0001234727,0.0003484713,0.0004209989,0.0002312122,0.00002123314,0.0005924577,0.0001882687,0.0002287396],"category_scores_gemma":[0.0004197843,0.00009403636,0.0001549037,0.001441326,0.001462397,0.00005801176,0.0001950733,0.001307554,0.000004154205],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00005204658,"about_ca_system_score_gemma":0.00009537689,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001533807,"about_ca_topic_score_gemma":0.0004260061,"domain_scores_codex":[0.9989167,0.0001014469,0.0002737207,0.0001924385,0.0003532042,0.0001624939],"domain_scores_gemma":[0.9973832,0.0002319245,0.0001259219,0.001474664,0.0006365927,0.0001476512],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"not_applicable","study_design_scores_codex":[0.001574379,0.005199467,0.3592491,0.001026744,0.02293464,0.0001246649,0.06710467,0.0000465247,0.1183411,0.07568961,0.3192023,0.02950674],"study_design_scores_gemma":[0.003765616,0.001031859,0.06303222,0.0008407474,0.006672798,0.0001370477,0.004084778,0.04277809,0.002012784,0.0001563284,0.8749477,0.0005400165],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"commentary","genre_gemma":"empirical","genre_scores_codex":[0.3016647,0.0110025,0.1329489,0.4631781,0.0005793759,0.001636502,0.0004329199,0.0007871509,0.08776993],"genre_scores_gemma":[0.9228439,0.0001116863,0.07021758,0.005962081,0.0000746783,0.0000143717,0.0007206883,0.00001881003,0.00003617159],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.6211793,"threshold_uncertainty_score":0.5680745,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.013589213883682,"score_gpt":0.3453327389953813,"score_spread":0.3317435251116993,"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."}}