{"id":"W4416962035","doi":"10.1109/pst65910.2025.11268849","title":"Harnessing Language Models to Analyze Android App Permission Fidelity","year":2025,"lang":"","type":"article","venue":"","topic":"Advanced Malware Detection Techniques","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Calgary","funders":"National Science Foundation","keywords":"Permission; Inference; Language model; Leverage (statistics); Transparency (behavior); Unpacking; Natural language; Fidelity; Categorization; Android (operating system)","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":[0.001390811,0.0008455156,0.0004453303,0.00151999,0.0003932159,0.001272549,0.000698243,0.0007773616,0.001459015],"category_scores_gemma":[0.01290398,0.0003819152,0.0006865567,0.0007544431,0.0007145072,0.003329291,0.001434201,0.001793384,0.001379259],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006767552,"about_ca_system_score_gemma":0.001149184,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01128521,"about_ca_topic_score_gemma":0.0181107,"domain_scores_codex":[0.9987331,0.0004072551,0.00008255203,0.0002991224,0.0003581594,0.0001199401],"domain_scores_gemma":[0.9943187,0.003731012,0.0003538981,0.0007558651,0.0007231617,0.0001174361],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0008036794,0.0005614863,0.0643678,0.0008518482,0.0002358999,0.001504827,0.00189179,0.3427754,0.04384363,0.01597586,0.02981347,0.4973742],"study_design_scores_gemma":[0.000009804999,0.00005303192,0.002988943,0.00002283586,0.00002141633,0.0001690626,0.0001512304,0.9792931,0.007019744,0.007447344,0.00279587,0.0000276785],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5140179,0.001897008,0.4444149,0.00205687,0.0002328671,0.0002380449,0.00745211,0.02174316,0.007947245],"genre_scores_gemma":[0.9310163,0.0003703625,0.0586885,0.0002756845,0.00005577151,0.0001189862,0.006076204,0.0005563966,0.002841685],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.01128521,"threshold_uncertainty_score":0.02243906,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0180555411407075,"score_gpt":0.3129693879139132,"score_spread":0.2949138467732056,"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."}}