CEDAR

A novel approach for silent CFE detection based on heterogeneous relation learning, which detects silent CFEs by localizing and hardening vulnerable instructions.

Runtime Environment：

Ubuntu 16.04 with 4.15 kernel；

Pytorch 1.10.2.

Usage:

Execute the scripts in the "Vulnerable instruction localization" folder to train on node classification task:

python3 main.py