Few-shot Website Fingerprinting attack with Meta-Bias Learning

•We investigate the under-studied, more realistic, and more challenging few-shot website fingerprinting attackproblem. Crucially, we focus on the model learning scalability for knowledge transfer efficacy, and model optimization strategy for task adaptation capability.•We propose a novel Meta-Bias Learning (MBL) method for solving few-shot WF attack. Specifically, we introduce a notion of parameter factorization, which avoids the need of meta-training the whole model. With this design, a majority fraction of parameters can be allocated to learn generic re-usable feature representations useful for all different tasks, whilst the remaining are used for more effective task adaptation.•Extensive experiments show that our MBL outperforms significantly previous state-of-the-art methods in both closed-world and open-world few-shot WF attack scenarios, with and without defense. Website fingerprinting (WF) attack aims to identify which website a user is visiting from the traffic data patterns. Whilst existing methods assume many training samples, we investigate a more realistic and scalable few-shot WF attack with only a few labeled training samples per website. To solve this problem, we introduce a novel Meta-Bias Learning (MBL) method for few-shot WF learning. Taking the meta-learning strategy, MBL simulates and optimizes the target tasks. Moreover, a new model parameter factorization idea is introduced for facilitating meta-training with superior task adaptation. Expensive experiments show that our MBL outperforms significantly existing hand-crafted feature and deep learning based alternatives in both closed-world and open-world attack scenarios, at the absence and presence of defense.