{"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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002750061,0.0002394511,0.0003073811,0.0001213399,0.0008122759,0.0003351327,0.001832061,0.00003412679,0.00002028819],"category_scores_gemma":[0.00002890348,0.0001980472,0.00001844697,0.000496898,0.00008949755,0.001172013,0.0004308057,0.0002040414,0.00001639565],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001194811,"about_ca_system_score_gemma":0.0001758978,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001779799,"about_ca_topic_score_gemma":0.0005525174,"domain_scores_codex":[0.9977412,0.0001102639,0.0003848296,0.0009018463,0.0004611769,0.0004006972],"domain_scores_gemma":[0.9972761,0.0003591878,0.0002564984,0.001803604,0.000147131,0.0001575113],"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":[0.00009228107,0.0005487482,0.006635398,0.0002552361,0.0002175822,0.00005386486,0.03435001,0.8197378,0.002617924,0.1252365,0.00626532,0.003989368],"study_design_scores_gemma":[0.0002096797,0.0001575144,0.00009975537,0.0001255521,0.00002142815,0.0000896441,0.005087747,0.9823343,0.0004883041,0.0008090085,0.01015478,0.0004223528],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.007116789,0.001291697,0.9882406,0.001167146,0.00007849275,0.001289004,0.0001444739,0.0003045009,0.0003672786],"genre_scores_gemma":[0.9700019,0.00004009721,0.02621485,0.0001321981,0.00005344686,0.003252918,0.00008110264,0.00003091632,0.0001925312],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9628851,"threshold_uncertainty_score":0.8076129,"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."}}