{"id":"W4406946785","doi":"10.1109/lgrs.2025.3536005","title":"SHAP-Assisted Resilience Enhancement Against Adversarial Perturbations in Optical and SAR Image Classification","year":2025,"lang":"en","type":"article","venue":"IEEE Geoscience and Remote Sensing Letters","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Adversarial system; Resilience (materials science); Computer science; Image (mathematics); Contextual image classification; Remote sensing; Optical imaging; Synthetic aperture radar; Artificial intelligence; Computer vision; Pattern recognition (psychology); Geology; Optics; Physics","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":[0.001396079,0.0008467016,0.0006189163,0.0005460565,0.0003257074,0.0006345807,0.0009341741,0.0008287932,0.001124573],"category_scores_gemma":[0.005258003,0.0002257568,0.000499242,0.0003356761,0.001162331,0.00146774,0.001623291,0.001301456,0.0002489717],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007291985,"about_ca_system_score_gemma":0.0005183168,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0009546757,"about_ca_topic_score_gemma":0.001108335,"domain_scores_codex":[0.9994164,0.0001790999,0.00002775236,0.0001382021,0.000165061,0.0000736065],"domain_scores_gemma":[0.9980308,0.0009610584,0.0003435165,0.0003408745,0.0002310587,0.00009273898],"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":[0.0001647425,0.0000569298,0.00256223,0.00007078875,0.00006562386,0.0001954639,0.00009952937,0.8318784,0.0152275,0.0317122,0.002143025,0.1158237],"study_design_scores_gemma":[0.000003153365,0.0000332523,0.0002312525,0.000004355215,0.000005623509,0.000032692,0.000007541264,0.9889037,0.002292102,0.008211698,0.0002693487,0.00000532384],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1054893,0.0003686989,0.8902789,0.0005326615,0.00006683866,0.00004431568,0.00008180993,0.000668276,0.002469272],"genre_scores_gemma":[0.9517621,0.0001759281,0.04618489,0.0001519473,0.00005605181,0.00002859059,0.000113141,0.00004250912,0.001484939],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.001396079,"threshold_uncertainty_score":0.007383227,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01215064829716259,"score_gpt":0.2638341849958837,"score_spread":0.2516835366987211,"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."}}