ICSE 2026 – Rio de Janeiro

We have presented our Journal-First paper, “Evaluating and Improving the Robustness of Security Attack Detectors Generated by LLMs,” at ICSE 2026.

---

In this work, we explore how Large Language Models can generate security-critical code, such as attack detectors, and how to make them more reliable. By combining Retrieval-Augmented Generation (RAG) with a Self-Ranking mechanism, we significantly improve detection performance for XSS and SQLi attacks.

I was also part of the organizing committee of the DeepTest 2026 workshop, bringing together researchers and practitioners working on testing AI systems and applying AI to software testing.

Great discussions and an amazing experience in Rio!

You can find some snapshots from this Journey in Moments