ResAware: Cross-Environment Website Fingerprinting via Resource-Privileged Distillation
Researchers have developed ResAware, a novel framework designed to improve the accuracy of website fingerprinting (WF) attacks across different environments. This framework utilizes a resource-aware distillation process, where a teacher model trained on resource-level features transfers its knowledge to a student model that operates using only encrypted traffic. Evaluations on a large dataset showed that ResAware significantly enhances the robustness of WF baselines, improving the F1-score of Var-CNN from 72.77% to 81.49% under a 150-day temporal drift. AI