VulBERTa: Simplified Source Code Pre-Training for Vulnerability Detection

VulBERTa architecture

Overview

This paper presents VulBERTa, a deep learning approach to detect security vulnerabilities in source code. Our approach pre-trains a RoBERTa model with a custom tokenisation pipeline on real-world code from open-source C/C++ projects. The model learns a deep knowledge representation of the code syntax and semantics, which we leverage to train vulnerability detection classifiers. We evaluate our approach on binary and multi-class vulnerability detection tasks across several datasets (Vuldeepecker, Draper, REVEAL and muVuldeepecker) and benchmarks (CodeXGLUE and D2A). The evaluation results show that VulBERTa achieves state-of-the-art performance and outperforms existing approaches across different datasets, despite its conceptual simplicity, and limited cost in terms of size of training data and number of model parameters.

Data

We provide all data required by VulBERTa.
This includes:

Tokenizer training data
Pre-training data
Fine-tuning data

Please refer to the data directory for further instructions and details.

Models

We provide all models pre-trained and fine-tuned by VulBERTa.
This includes:

Trained tokenisers
Pre-trained VulBERTa model (core representation knowledge)
Fine-tuned VulBERTa-MLP and VulBERTa-CNN models

Please refer to the models directory for further instructions and details.

Pre-requisites and requirements

In general, we used this version of packages when running the experiments:

Python 3.8.5
Pytorch 1.7.0
Transformers 4.4.1
Tokenizers 0.10.1
Libclang (any version > 12.0 should work. https://pypi.org/project/libclang/)

For an exhaustive list of all the packages, please refer to requirements.txt file.

How to use

In our project, we uses Jupyterlab notebook to run experiments.
Therefore, we separate each task into different notebook:

Pretraining_VulBERTa.ipynb - Pre-trains the core VulBERTa knowledge representation model using DrapGH dataset.
Finetuning_VulBERTa-MLP.ipynb - Fine-tunes the VulBERTa-MLP model on a specific vulnerability detection dataset.
Evaluation_VulBERTa-MLP.ipynb - Evaluates the fine-tuned VulBERTa-MLP models on testing set of a specific vulnerability detection dataset.
Finetuning+evaluation_VulBERTa-CNN - Fine-tunes VulBERTa-CNN models and evaluates it on a testing set of a specific vulnerability detection dataset.

Running VulBERTa-CNN or VulBERTa-MLP on arbitrary codes

Coming soon!

Citation

Accepted as conference paper (oral presentation) at the International Joint Conference on Neural Networks (IJCNN) 2022.
Link to paper: https://ieeexplore.ieee.org/document/9892280

@INPROCEEDINGS{hanif2022vulberta,

  author={Hanif, Hazim and Maffeis, Sergio},
  booktitle={2022 International Joint Conference on Neural Networks (IJCNN)}, 
  title={VulBERTa: Simplified Source Code Pre-Training for Vulnerability Detection}, 
  year={2022},
  volume={},
  number={},
  pages={1-8},
  doi={10.1109/IJCNN55064.2022.9892280}

}

ICL-ml4csec / VulBERTa

readme