{"id":"W4220681850","doi":"10.1016/j.compbiomed.2022.105341","title":"Collaborative learning of weakly-supervised domain adaptation for diabetic retinopathy grading on retinal images","year":2022,"lang":"en","type":"article","venue":"Computers in Biology and Medicine","topic":"Retinal Imaging and Analysis","field":"Medicine","cited_by":37,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Alberta","funders":"","keywords":"Artificial intelligence; Computer science; Discriminator; Domain adaptation; Diabetic retinopathy; Pattern recognition (psychology); Machine learning; Binary classification; Grading (engineering); Support vector machine; Classifier (UML); Medicine; Detector","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":[0.0007649852,0.0001068695,0.0004175414,0.0002750409,0.0001439081,0.000002001552,0.00005259903,0.00003568355,0.00001691013],"category_scores_gemma":[0.0002519593,0.00008570177,0.00003854636,0.0003362302,0.0002668269,0.00001461543,0.00003764721,0.0002534498,2.189079e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003998727,"about_ca_system_score_gemma":0.00003185236,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001772133,"about_ca_topic_score_gemma":5.754845e-7,"domain_scores_codex":[0.998966,0.0002597791,0.0002709842,0.000250047,0.00009533581,0.0001578786],"domain_scores_gemma":[0.9991777,0.0004780661,0.0001268021,0.00009076831,0.00007505383,0.00005161236],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.005037094,0.0004595146,0.6599219,0.000649315,0.0003634634,0.0001510673,0.02279734,0.001367399,0.2169454,0.01308592,0.00237836,0.0768432],"study_design_scores_gemma":[0.04680843,0.06819136,0.429983,0.005335778,0.001524989,0.0004858671,0.0979474,0.2846181,0.01101298,0.0325051,0.02003718,0.001549718],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9796352,0.001185145,0.01205527,0.00610057,0.0002313534,0.0002940518,0.000007168633,0.00002179746,0.000469465],"genre_scores_gemma":[0.9920711,0.000138263,0.006991093,0.0004694237,0.0001035824,0.00003701582,0.0001021686,0.000008328213,0.00007901889],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2832507,"threshold_uncertainty_score":0.3494817,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01391377109063835,"score_gpt":0.3072750645403019,"score_spread":0.2933612934496636,"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."}}