Bartolomeo Stellato
Papers
5
Total Citations
86
H-Index
5
About
Bartolomeo Stellato is a leading researcher at the intersection of optimization, machine learning, and robotics. His work focuses on making complex, mixed-integer convex programming (MICP) fast enough for real-time robotic control—a challenge that has long limited the application of these powerful models in the physical world. Stellato’s major contributions include developing supervised learning frameworks that dramatically accelerate MICP solvers, enabling online motion planning and manipulation that were previously computationally infeasible. His highly cited paper "CoCo: Online Mixed-Integer Control Via Supervised Learning" (2021, 36 citations) exemplifies this breakthrough, while his foundational work "Learning Mixed-Integer Convex Optimization Strategies for Robot Planning and Control" (2020, 25 citations) has shaped how researchers approach real-world robotic decision-making. Stellato has also explored the theoretical underpinnings of continuous control, investigating why trained agents gravitate toward boundary actions in reinforcement learning. His recent work on Stackelberg games for multi-robot coordination further demonstrates his commitment to scalable, principled solutions. With over 85 citations across his top papers, Stellato is a rising voice in optimization-driven robotics, bridging the gap between theoretical rigor and practical deployment.
Research Focus
Key Achievements
Top Papers
- 1CoCo: Online Mixed-Integer Control Via Supervised Learning36 citations · 2021
- 2
- 3
- 4
- 5