{"id":"W4408549997","doi":"10.1029/2024wr037398","title":"On Robustness of the Explanatory Power of Machine Learning Models: Insights From a New Explainable AI Approach Using Sensitivity Analysis","year":2025,"lang":"en","type":"article","venue":"Water Resources Research","topic":"Explainable Artificial Intelligence (XAI)","field":"Computer Science","cited_by":23,"is_retracted":false,"has_abstract":true,"ca_institutions":"Global Institute for Water Security; University of Saskatchewan","funders":"Natural Sciences and Engineering Research Council of Canada; Suncor Energy Incorporated","keywords":"Robustness (evolution); Explanatory power; Sensitivity (control systems); Machine learning; Artificial intelligence; Computer science; Engineering; Epistemology; Chemistry","routes":{"ca_aff":true,"ca_fund":true,"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":[0.002095323,0.0002047354,0.000471677,0.001211565,0.0004966727,0.0001861702,0.001452435,0.0001367852,0.00002459437],"category_scores_gemma":[0.0002413359,0.0001307695,0.0002225654,0.003089529,0.0002891777,0.0004740678,0.001435278,0.0007873885,0.000005059561],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001486687,"about_ca_system_score_gemma":0.0001449032,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01563455,"about_ca_topic_score_gemma":0.0004199354,"domain_scores_codex":[0.9949241,0.002059482,0.0004769122,0.0006955101,0.001239943,0.0006040537],"domain_scores_gemma":[0.9973028,0.0006807551,0.0001127461,0.001320561,0.0004700684,0.0001131158],"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":[0.00009253478,0.0001655516,0.0007480548,0.00004219584,0.0002462409,0.00001432699,0.01784332,0.9630841,0.01050665,0.006650103,0.00006298228,0.0005439594],"study_design_scores_gemma":[0.0001066191,0.00004453733,0.000106827,0.00007063452,0.00003090683,7.88491e-7,0.00136305,0.8428431,0.1468792,0.008143902,0.0002931069,0.0001173806],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6273084,0.0001937527,0.3705611,0.000210089,0.00004697049,0.0002051876,0.000002808502,0.00002813168,0.001443476],"genre_scores_gemma":[0.9963413,0.000006312014,0.002340153,0.00004902894,0.00002610374,0.00000955315,0.000005277216,0.00001421495,0.00120803],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3690329,"threshold_uncertainty_score":0.9909204,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.08346191017698525,"score_gpt":0.318397046788249,"score_spread":0.2349351366112638,"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."}}