Uploaded November 2025 | Updated September 2026, 4 hours ago
Traditional robot learning in imitation and reinforcement learning (IL/RL) often predicts per-step actions or short action segments, which can yield discontinuous trajectories, weak exploration, and high inference cost. Movement Primitives (MPs) provide a compact, parameterized trajectory representation that captures shape and inter-segment correlations in a low-dimensional latent space. In this talk, Ge Li presents MPs as an efficient action-sequence model and shows how to integrate them with modern IL/RL pipelines. The resulting systems produce smoother trajectories, explore more effectively in RL settings, and better encode motion correlations, while reducing inference and representation cost— including in vision-language models.
Ge Li is a PhD candidate at Karlsruhe Institute of Technology (KIT), Germany. His research focuses on reinforcement learning and continuous control for robot manipulation. He holds an M.Sc. from RWTH Aachen University and a B.Sc. from the University of Science and Technology of China. Before his PhD, he was a research intern at the Max Planck Institute for Intelligent Systems and a robot software engineering intern at Kärcher.
Traditional robot learning in imitation and reinforcement learning (IL/RL) often predicts per-step actions or short action segments, which can yield discontinuous trajectories, weak exploration, and high inference cost. Movement Primitives (MPs) provide a compact, parameterized trajectory representation that captures shape and inter-segment correlations in a low-dimensional latent space. In this talk, Ge Li presents MPs as an efficient action-sequence model and shows how to integrate them with modern IL/RL pipelines. The resulting systems produce smoother trajectories, explore more effectively in RL settings, and better encode motion correlations, while reducing inference and representation cost— including in vision-language models.
Ge Li is a PhD candidate at Karlsruhe Institute of Technology (KIT), Germany. His research focuses on reinforcement learning and continuous control for robot manipulation. He holds an M.Sc. from RWTH Aachen University and a B.Sc. from the University of Science and Technology of China. Before his PhD, he was a research intern at the Max Planck Institute for Intelligent Systems and a robot software engineering intern at Kärcher.
