Max Robotics — Global Service Robot Directory + US Certification

About

Giulio Mazzi is an AI researcher specializing in safe and explainable decision-making under uncertainty, with a particular focus on Partially Observable Markov Decision Processes (POMDPs) and the Partially Observable Monte Carlo Planning (POMCP) algorithm. His work addresses one of the most pressing challenges in deploying autonomous systems: ensuring that powerful but opaque planning algorithms behave safely and predictably in real-world environments. Mazzi's most significant contributions center on developing shielding mechanisms for POMCP policies — techniques that detect and prevent potentially unsafe or unexpected decisions at runtime. His 2023 paper on risk-aware shielding (12 citations) represents his most impactful work, extending earlier rule-based approaches to account for uncertainty in partially observable settings. Complementary research on identifying unexpected decisions and explaining the influence of prior knowledge on POMCP policies demonstrates his commitment to making these algorithms both safer and more interpretable. Collectively accumulating over 30 citations, his body of work has meaningfully advanced the intersection of formal verification and online planning, offering practical tools for researchers and practitioners seeking to build trustworthy AI systems capable of operating reliably in complex, uncertain environments.