Automated Cross-Repository Vulnerability Variant Retrieval Using Patch-Weighted Contrastive Learning
In Plain Terms
When a security flaw is found in one open-source project, copies and near-copies of that same flawed code often live on in other projects. This paper presents AVA, a system that learns from known vulnerabilities and their fixes to automatically hunt down those unfixed look-alikes across many repositories at once. It shows that real-world patches can teach a model to tell vulnerable code from fixed code, extending remediation beyond the project where a flaw was first reported.
Artifacts
Related Papers
Cite this
APA Chen, J., Ampel, B. M., Ullman, S., Reyes, R. Y., & Chen, H. (2026). Automated Cross-Repository Vulnerability Variant Retrieval Using Patch-Weighted Contrastive Learning. In Proceedings of the 2026 IEEE Cyber Awareness and Research Symposium (CARS).
BibTeX
@inproceedings{chen2026automated,
title={Automated Cross-Repository Vulnerability Variant Retrieval Using Patch-Weighted Contrastive Learning},
author={Joseph Chen and Benjamin M. Ampel and Steven Ullman and Raul Y. Reyes and Hsinchun Chen},
booktitle={Proceedings of the 2026 IEEE Cyber Awareness and Research Symposium (CARS)},
year={2026}
}
Benjamin M. Ampel
Assistant Professor in Computer Information Systems and Director, Center for CyberAI Research (CCAIR)
My research focuses on AI-enabled Cybersecurity, including Cyber Threat Intelligence, Large Language Models, and Phishing Detection.