Authors: Davy Preuveneers / KUL and Valentina Del Rio / PLURIBUS
We are excited to announce that our paper, “The Adaptive Arms Race: Redefining Robustness in AI Security”, has been accepted for presentation at the 28th International Symposium on Research in Attacks, Intrusions and Defenses (RAID 2025), which will take place on 19–22 October 2025 in Gold Coast, Australia.
This work is the result of a great collaboration between Ilias Tsingenopoulos, Vera Rimmer, Fabio Pierazzi, Lorenzo Cavallaro, and Wouter Joosen. It explores the dynamic arms race between adversaries and defenders in AI security, and how we should rethink robustness when facing adaptive threats.
RAID is one of the top-tier cybersecurity conferences, and we are proud to continue our presence there after last year’s publication on reinforcement learning–based antivirus hardening.
While the final version will be published later this year by IEEE, an earlier version is available on arXiv for those interested in a preview (note: the accepted version includes important updates and acknowledgements).
We are looking forward to sharing more soon
Stay tuned!