BinGNN

Binary Security Patch Identification with Graph Neural Networks.

Dependencies

conda create -n pytorch_cu117 python=3.9
conda install pytorch torchvision torchaudio pytorch-cuda=11.7 -c pytorch -c nvidia
conda install pyg -c pyg
conda install pandas matplotlib networkx natsort

Data Preprocessing

Model Training

Basic Analysis

All operations are performed under the root path.

1. Plot Training Curve.

python analysis/plot_train_curve.py -log <path-to-training-log>.txt

The results are saved in the same path and the same filename with .png.

2. Plot Graph Structure.

python analysis/plot_graph.py -npz <path-to-npz-graph-file>.npz

The results are saved in the ./temp/ with a .txt file (records) and a .png file (visualization).

3. Plot Graph Statistic.

python analysis/plot_graph_dist.py

The results are saved in the ./temp/records.csv and ./temp/records.png.