{"id":"W4387640457","doi":"10.1007/978-3-031-45676-3_32","title":"MEDKD: Enhancing Medical Image Classification with Multiple Expert Decoupled Knowledge Distillation for Long-Tail Data","year":2023,"lang":"en","type":"book-chapter","venue":"Lecture notes in computer science","topic":"Imbalanced Data Classification Techniques","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of British Columbia","funders":"","keywords":"Computer science; Contextual image classification; Class (philosophy); Machine learning; Artificial intelligence; Distillation; Image (mathematics); Task (project management); Data mining; Noise (video); Pattern recognition (psychology)","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":["metaepi_narrow","open_science"],"consensus_categories":[],"category_scores_codex":[0.002346461,0.0005760071,0.000583442,0.0008089393,0.0003831714,0.0006578565,0.007388976,0.0004793725,0.00001495907],"category_scores_gemma":[0.001502992,0.0004957432,0.0000755449,0.0009187296,0.0008112906,0.001650181,0.002330985,0.000646762,0.00006603036],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004452934,"about_ca_system_score_gemma":0.001385691,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002123627,"about_ca_topic_score_gemma":0.001469142,"domain_scores_codex":[0.9942622,0.00005357031,0.0008450967,0.002652165,0.001480586,0.0007064099],"domain_scores_gemma":[0.9924574,0.002073972,0.0005495282,0.004011832,0.0006456611,0.0002615289],"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":[0.00003109683,0.00008591709,0.0001925028,0.0001689656,0.00002890998,0.0000377218,0.000732242,0.0002823857,0.001584461,0.01622788,0.000752299,0.9798756],"study_design_scores_gemma":[0.0004475878,0.0001154579,0.0008063503,0.0007256056,0.000009700978,0.00003174738,6.450359e-7,0.98014,0.003558191,0.01085731,0.002631212,0.0006762499],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.00001644222,0.0001778628,0.9945968,0.001722208,0.001050211,0.001102665,0.00007372295,0.0008643282,0.000395735],"genre_scores_gemma":[0.05560696,0.0001848043,0.9413133,0.0004798267,0.0007947712,0.0001847755,0.0008958126,0.0001158535,0.0004238711],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.9798576,"threshold_uncertainty_score":0.9997494,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05536565858843196,"score_gpt":0.325771751090763,"score_spread":0.2704060925023311,"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."}}