Pro-ZD: A Transferable Graph Neural Network Approach for Proactive Zero-Day Threats Mitigation
Nardine Basta § , Firas Ben Hmida § , Houssem Jmal § , Muhammad Ikram, Mohamed Ali Kaafar, and Andy Walker · 2026
Read Pro-ZD: A Transferable Graph Neural Network Approach for Proactive Zero-Day Threats Mitigation by Nardine Basta § , Firas Ben Hmida § , Houssem Jmal § , Muhammad Ikram, Mohamed Ali Kaafar, and Andy Walker (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-07073.txt.
Description
Nardine Basta and five co-authors build a proactive defense for enterprise networks against zero-day attacks. In today's firms, perimeter plus distributed firewall rules govern connectivity. Automated tools now generate those rules and access policies, yet risks from dynamically generated policies, above all exposure of critical assets, remain hard to manage. Networks keep shifting too, with remote users, personal devices, and cloud links.
Their system, Pro-ZD, uses a graph neural network to find weighted shortest paths through the network. Those paths reveal misconfigurations and high-risk links that could threaten critical assets and be exploited through undisclosed flaws. Pro-ZD then tunes firewall rules and access policies on its own to close high-risk links and block unauthorized access. Tests show robustness and transferability, with over 95 percent average accuracy at spotting high-risk links across settings.
Catalog record
| CREATOR | Nardine Basta § , Firas Ben Hmida § , Houssem Jmal § , Muhammad Ikram, Mohamed Ali Kaafar, and Andy Walker |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 71 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-07073