{"id":"W4310496924","doi":"10.1016/j.artmed.2022.102471","title":"Why did AI get this one wrong? — Tree-based explanations of machine learning model predictions","year":2022,"lang":"en","type":"article","venue":"Artificial Intelligence in Medicine","topic":"Explainable Artificial Intelligence (XAI)","field":"Computer Science","cited_by":52,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Ottawa","funders":"European Commission","keywords":"Computer science; Machine learning; Artificial intelligence; Boosting (machine learning); Decision tree; Ensemble learning; Fidelity; Gradient boosting; Random forest","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","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.002118949,0.0003233361,0.0005642329,0.001139155,0.000709875,0.00005981351,0.001715769,0.00009942942,0.001244221],"category_scores_gemma":[0.0009882866,0.0003402552,0.0001291145,0.002926865,0.0004366392,0.0006302689,0.0004358202,0.001113013,0.00006082882],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003148075,"about_ca_system_score_gemma":0.0002961847,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003717911,"about_ca_topic_score_gemma":0.002664636,"domain_scores_codex":[0.9952797,0.000462445,0.00148797,0.0008106921,0.001281559,0.0006776294],"domain_scores_gemma":[0.9973846,0.0007729487,0.0003848353,0.0009077933,0.000346142,0.0002036554],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0000744737,0.0006784661,0.0008952826,0.00002622225,0.0000240582,0.00003426056,0.005343143,0.7889012,0.005760436,0.1678803,0.001078607,0.02930362],"study_design_scores_gemma":[0.00008088402,0.0006729873,0.00003241037,0.0000733606,0.00002321333,0.000006229764,0.001462183,0.9126665,0.0204702,0.06238968,0.001849349,0.0002730143],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.005146772,0.0003461955,0.9688201,0.02162376,0.000610235,0.0005248203,0.00002132534,0.0003489334,0.002557867],"genre_scores_gemma":[0.9850253,0.00006747036,0.01125816,0.002928738,0.0001291291,0.0002637105,0.00004164024,0.0000438138,0.0002420209],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9798785,"threshold_uncertainty_score":0.9999049,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0677420181985633,"score_gpt":0.3151730733071335,"score_spread":0.2474310551085702,"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."}}