{"id":"W3112198773","doi":"10.1016/j.eswa.2022.117230","title":"When stakes are high: Balancing accuracy and transparency with Model-Agnostic Interpretable Data-driven suRRogates","year":2022,"lang":"en","type":"article","venue":"Expert Systems with Applications","topic":"Explainable Artificial Intelligence (XAI)","field":"Computer Science","cited_by":26,"is_retracted":false,"has_abstract":false,"ca_institutions":"Université Laval","funders":"Fonds Wetenschappelijk Onderzoek","keywords":"Categorical variable; Computer science; Feature selection; Black box; Surrogate model; Feature engineering; Generalized linear model; Transparency (behavior); Decision tree; Machine learning; Segmentation; Gradient boosting; Artificial intelligence; Data mining; Boosting (machine learning); Variable (mathematics); Random forest; Mathematics; Deep learning","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.05699833,0.001127021,0.0024023,0.001984165,0.001386394,0.01274968,0.003772703,0.005628845,0.003343357],"category_scores_gemma":[0.3269694,0.00130824,0.001187534,0.00165883,0.007379781,0.01882524,0.008196112,0.008655488,0.0006466124],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.003820852,"about_ca_system_score_gemma":0.004266196,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001825542,"about_ca_topic_score_gemma":0.001762991,"domain_scores_codex":[0.9282557,0.04916067,0.002616131,0.004957727,0.01232785,0.002681943],"domain_scores_gemma":[0.6562676,0.2652084,0.02378012,0.04041129,0.01070937,0.003623195],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0009077761,0.0001813147,0.007568332,0.0002917748,0.0001881434,0.0003341537,0.001571662,0.222727,0.001338814,0.7117702,0.002664095,0.05045673],"study_design_scores_gemma":[0.00006189976,0.00007049509,0.0008420289,0.0001365955,0.00004240819,0.00007138574,0.000198593,0.2970448,0.001237506,0.6985498,0.001695473,0.00004904039],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.09886368,0.000763764,0.8628996,0.01741634,0.0002879801,0.000196039,0.0005392121,0.0007549875,0.01827829],"genre_scores_gemma":[0.9340578,0.0001422121,0.06380653,0.0005603416,0.00009626698,0.0001084487,0.0001470491,0.0001479537,0.000933299],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.05699833,"threshold_uncertainty_score":0.3014396,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04006494069805255,"score_gpt":0.2687050028251239,"score_spread":0.2286400621270714,"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."}}