Published January 17, 2023 | Version v4
Software Open

VULGEN: Realistic Vulnerability Generation Via Pattern Mining and Deep Learning

Creators

  • 1. Washington State University

Description

By combining the strengths of state-of-the-art deterministic (pattern-based) approach for vulnerability injection and probabilisitic (deep-learning/DL-based) program transformation approach for injection localization, we present VulGen, which is the first injection-based vulnerability-generation technique that is not limited to a particular class of vulnerabilities. We compare VulGen with several other possible techniques (T5, Graph2Edit, Getafix) for vulnerability generation and show that VulGen outperforms them.

Once the users have Docker installed download the Docker image "vulgen_image.tar.xz"

Then, check the README.md for detailed steps of reproducing the experiments.

Besides, we also provide the simple package of the artifact "vulgen.zip". The raw data of our experiments is also provided in this simple package. However, using it to reproduce the experiments requires the users to set up the enviroments and dependencies, which is not recommanded.

 

Files

Documents_for_AE.zip

Files (18.1 GB)

Name Size Download all
md5:f345fc46e6b3cebfe6b19e07ff47a87d
225.6 kB Preview Download
md5:58e0f97b7afac84d4ff0eefeca5ba711
150 Bytes Download
md5:836b15976313682389dd3d9e32bd9a1b
10.7 kB Preview Download
md5:9b1b18ef5a593d551f6d660ee2f90322
6.8 GB Preview Download
md5:98b3e0a277e608d515a50da9373eb247
11.2 GB Download