Thibaut Kulak

About

Thibaut Kulak is a roboticist focused on making robot programming accessible to non-experts through intuitive, data-driven methods. His key research areas lie at the intersection of Learning from Demonstration (LfD) and robot skill acquisition, with a particular emphasis on probabilistic movement primitives (ProMPs) and intrinsically motivated learning. Kulak’s major contributions include the development of "Fourier movement primitives" (FMPs), a novel representation for learning rhythmic, periodic skills—such as those needed in factory or household tasks—from human demonstrations. His work on "Active Learning of Bayesian Probabilistic Movement Primitives" (19 citations) advances the field by enabling robots to actively query a teacher for the most informative demonstrations, significantly improving learning efficiency. Additionally, his research on combining social guidance with intrinsically motivated learning (2 citations) tackles the challenge of multitask skill acquisition, allowing robots to autonomously explore and refine a repertoire of movements. Through these contributions, Kulak is helping to bridge the gap between human intuition and robotic dexterity, paving the way for more adaptable and user-friendly robotic systems in real-world environments.

Research Focus

Key Achievements

Top Papers

1. 1
   Active Learning of Bayesian Probabilistic Movement Primitives

   19 citations · 2021
2. 2
   Fourier movement primitives: an approach for learning rhythmic robot skills from demonstrations

   15 citations · 2020
3. 3
   Combining Social and Intrinsically Motivated Learning for Multitask Robot Skill Acquisition

   2 citations · 2021

Key Collaborators

Contact & Links