Triplet Fingerprinting: More Practical and Portable Website Fingerprinting with N-shot Learning
Published in ACM SIGSAC Conference on Computer and Communications Security (CCS), 2019
In this work, we propose a new website fingerprinting (WF) attack called Triplet Fingerprinting (TF) that leverages triplet networks and N-shot learning to reduce the need for large amounts of training data. The TF attack remains effective in challenging settings such as those involving different network conditions and when training and testing data are collected years apart. Notably, our approach can achieve up to 85% accuracy with only five examples per website. This makes it highly practical for attackers with limited data and computational resources, representing a significant step forward in the portability and practicality of WF attacks.
Recommended citation: Payap Sirinam, Nate Mathews, Mohammad Saidur Rahman, Matthew Wright. (2019). "Triplet Fingerprinting: More Practical and Portable Website Fingerprinting with N-shot Learning." ACM SIGSAC Conference on Computer and Communications Security (CCS), 2019. DOI: 10.1145/3319535.3354217.
Download Paper | Download Slides