Alberto Dalla Libera
Papers
9
Total Citations
72
H-Index
5
About
Alberto Dalla Libera is a robotics and machine learning researcher whose work sits at the intersection of data-driven modeling, dynamics identification, and reinforcement learning for physical systems. His research focuses on developing intelligent algorithms that allow robots to learn and adapt from data, particularly in scenarios where traditional model-based approaches fall short due to limited system knowledge or constrained sensing capabilities. Among his most significant contributions is a physics-informed Gaussian Process (GP) framework for robot inverse dynamics identification, which elegantly embeds physical laws into black-box estimators to improve generalization and data efficiency — his most cited work with 20 citations. Complementing this, his derivative-free model learning framework for reinforcement learning addresses the practical challenge of systems where only positions are measurable, earning 16 citations. He has also made notable strides in autonomous kinematic model learning (14 citations) and proprioceptive collision detection using GP regression. More recently, Dalla Libera has expanded into model-based reinforcement learning with large language model integration, reflecting a forward-looking vision for autonomous robotic manipulation. Across his portfolio, his consistent use of Gaussian Process Regression as a principled probabilistic tool underscores a commitment to uncertainty-aware, physically grounded robot learning — making his work highly relevant for researchers advancing autonomous and adaptable robotic systems.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3Autonomous Learning of the Robot Kinematic Model14 citations · 2020
- 4
- 5Forward Dynamics Estimation from Data-Driven Inverse Dynamics Learning??5 citations · 2023
- 6Learning Control from Raw Position Measurements4 citations · 2023
- 7
- 8
- 9