Publications

Aleksandar Fontana*, Marco Simoni*, Giulio Rossolini, Andrea Saracino, Paolo Mori (2026). On the Hidden Objective Biases of Group-based Reinforcement Learning. ACL.

Aleksandar Fontana, Marco Simoni (2025). Unmasking model behavior: How llms reason on vulnerability detection. International Conference on Availability, Reliability and Security.

Source Document

Marco Simoni, Aleksandar Fontana, Andrea Saracino, Paolo Mori (2025). TITAN: Graph-Executable Reasoning for Cyber Threat Intelligence. arXiv preprint arXiv:2510.14670.

Marco Simoni*, Aleksandar Fontana*, Giulio Rossolini, Andrea Saracino (2025). Improving LLM Reasoning for Vulnerability Detection via Group Relative Policy Optimization. arXiv preprint arXiv:2507.03051.

Marco Simoni*, Aleksandar Fontana*, Giulio Rossolini, Andrea Saracino, Paolo Mori (2025). GTPO: Stabilizing Group Relative Policy Optimization via Gradient and Entropy Control. arXiv preprint arXiv:2508.03772.