MIRROR

Xiaochuan Zhang, Wenjie Sun, Jianmin Pang, Fudong Liu and Zhen Ma

Table of Contents

1. Introduction
2. Citation
3. Installation
4. Requirements and Dependencies
5. Data Preprocessing
6. Pretrain
7. Train

Introduction

This is the the prototype system of paper Similarity Metric Method for Binary Basic Blocks of Cross-Instruction Set Architecture.

Citation

If you find the code and datasets useful in your research, please cite:

@inproceedings{mirror,
  title={Similarity Metric Method for Binary Basic Blocks of Cross-Instruction Set Architecture},
  author={Xiaochuan, Zhang and Wenjie, Sun and Jianmin, Pang and Fudong, Liu and Zhen, Ma},
  booktitle={Proceedings of the NDSS Workshop on Binary Analysis Research},
  year={2020}
}

Requirements and Dependencies

• Ubuntu (We test with Ubuntu = 18.04 LTS)
• Python (We test with Python = 3.7.4)
• CUDA & cuDNN (We test with CUDA = 10.1 and cuDNN = 7.6.5)
• PyTorch （We test with PyTorch = 1.0.0）
• NVIDIA GPU(s) (We use 4 RTX 2080Ti)

Installation

Download repository:

$ git clone https://github.com/zhangxiaochuan/MIRROR.git
$ cd MIRROR

The dataset MISA (Multi-ISAs basic block dataset) is available at link. Please download and uncompress MISA.zip, and place MISA in the root directory of the project.

Data Preprocessing

$ python data_manager.py

Pretrain

$ python pretrain.py

train

$ python train.py

Contact

Xiaochuan Zhang

License

See MIT License