Silvio Giancola

King Abdullah University of Science and Technology, Politecnico di Milano

Papers

Total Citations

H-Index

About

Silvio Giancola is a leading researcher at the intersection of 3D computer vision and trustworthy deep learning, with a primary focus on point cloud analysis. His work addresses two critical challenges: designing efficient architectures for 3D data and certifying the robustness of these models against real-world threats. Giancola’s most notable contribution is **LC-NAS**, a latency-constrained neural architecture search framework that automatically designs high-accuracy point cloud networks while respecting computational budgets—a vital step for deploying 3D models on edge devices. Complementing this, his work on **3DeformRS** pioneers a certified defense against spatial deformations, providing provable robustness guarantees for point cloud models used in safety-critical domains like autonomous driving and surgical robotics. With over 20 citations across his top papers, Giancola’s research is gaining traction for its practical impact. His earlier work on multi-camera trajectory programming for robotic painting further showcases his versatility in bridging perception and robotics. Giancola’s contributions are shaping a future where 3D vision systems are not only powerful but also reliable and efficient.