{"id":"W7125811146","doi":"10.21428/594757db.f2451ce9","title":"Faithful Perturbations and Evaluations for Post-Hoc Local Explanation Methods","year":2025,"lang":"en","type":"article","venue":"","topic":"Explainable Artificial Intelligence (XAI)","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Alberta","funders":"University of Alberta; Natural Sciences and Engineering Research Council of Canada; Canadian Institute for Advanced Research","keywords":"Fidelity; Noise (video); Perturbation (astronomy); Quality (philosophy); Distribution (mathematics); Worry","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":[0.0007913791,0.00009069167,0.0001007682,0.00021665,0.000314599,0.000188216,0.0003224182,0.00005547371,0.0000314256],"category_scores_gemma":[0.0005366018,0.00008341971,0.00004096306,0.0004803939,0.00005590635,0.0006049325,0.0001110081,0.00005491726,0.00001554988],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00006500743,"about_ca_system_score_gemma":0.0001533702,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00007202955,"about_ca_topic_score_gemma":0.0001681998,"domain_scores_codex":[0.9990327,0.0001131212,0.0002270669,0.0003162896,0.0001254952,0.0001853137],"domain_scores_gemma":[0.9982474,0.0008784651,0.00003928657,0.0003122577,0.0004711423,0.00005145406],"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":[0.000003982217,0.00001889464,0.000009842166,0.000006250527,0.000008153314,1.42235e-7,0.0004940331,0.0002775251,0.001172845,0.6302369,0.001137692,0.3666337],"study_design_scores_gemma":[0.0001195996,0.0001186212,0.0003875087,0.00001050127,0.00001486079,0.000002048986,0.001775251,0.8231276,0.02420695,0.1378017,0.01231284,0.0001225037],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.001165939,0.0001783357,0.9804355,0.01076771,0.0002611571,0.0004134388,0.000003102052,0.0001366311,0.006638175],"genre_scores_gemma":[0.1757975,0.00001771384,0.8173023,0.001891945,0.00002716858,0.0002200192,0.00001216475,0.000005857332,0.004725329],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.82285,"threshold_uncertainty_score":0.3401757,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03886993933859015,"score_gpt":0.4173731409240598,"score_spread":0.3785032015854697,"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."}}