Development of Advanced AI and Deep Learning Models for Security Patches
| Reference number | |
| Coordinator | Högskolan i Halmstad |
| Funding from Vinnova | SEK 145 000 |
| Project duration | October 2024 - March 2025 |
| Status | Ongoing |
| Venture | International individual mobility within cutting-edge technology |
| Call | Closed offer - International individual mobility for cutting-edge technology 2024 |
Important results from the project
The main objective of this project was to develop robust models capable of identifying and predicting vulnerabilities in open-source software. The researcher developed the vulnerabilities prediction system based on applying AI advanced technologies. In this collaboration, the project achieved the promising performance over the state-of-the-art. The results were published and presented in CSA 2024 conference in Thailand and ICIAI 2025 conference in Singapore.
Expected long term effects
The results of this project are expected to significantly enhance the reliability, transparency, and adoption of AI-driven vulnerability detection systems in software development. Over time, the integration of explainable AI techniques will support the widespread adoption of automated security solutions across large-scale software pipelines, ultimately contributing to more secure software ecosystems and increasing organizational resilience against evolving cybersecurity threats.
Approach and implementation
The project was implemented based on the planed proposal and the researcher from Halmstad University tasks at Korea University focused on the development of the novel security patches detection system through C/C++ vulnerabilities category. Researcher´s stay in South Korea allowed the successful implementation of the design process. The project followed the planed timeline and had a meaningful collaboration.