Debunking Represenation Learning for Encrypted Traffic Classification

This repository contains the implementation of various baseline algorithms and data preprocessing methods.

Folder Structure

• code: This folder contains the implementation of various baseline algorithms.
• process_finetune_data: This folder contains the methods for finetuned data preprocessing.

Baseline Algorithms

The code folder includes the following baseline algorithms:

• Pcap Encoder
• ET-BERT
• YaTC
• NetMamba
• TrafficFormer
• netFound
• AutoGluon based Feature Engineering

Many thanks to authors for contributions.

Data Preprocessing

The process_finetune_data folder includes the following data preprocessing methods:

• Split: Per-packet/Per-flow split
• Filter: How to filter the unrelevant protocal packets
• Data processing (pkt/flow) of different models

Data is available on HuggingFace

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• Thanks to all previous works.