Takayuki Osa
Takayuki Osa develops algorithms that let robots learn behavior and adapt to changing tasks. His research combines reinforcement learning, imitation and motion planning. Osa, Naohiko Sugita and Mamoru Mitsuishi coauthored work on online trajectory planning in dynamic environments for surgical task automation. The work appeared at RSS 2014. Osa and five colleagues coauthored An Algorithmic Perspective on Imitation Learning. The work surveys methods for learning actions from demonstrations.
Timeline
-
2014
Planning surgical robot motion
Osa, Naohiko Sugita and Mamoru Mitsuishi coauthored work on online trajectory planning in dynamic environments for surgical task automation. The work appeared at RSS 2014.
-
2018
A view of imitation learning
Osa and five colleagues coauthored An Algorithmic Perspective on Imitation Learning. The work surveys methods for learning actions from demonstrations.
-
2022
Learning a space of motion solutions
Osa published work on motion planning by learning the solution manifold in trajectory optimization. This extends his earlier research on finding multiple trajectories for a task.
-
2024
Multiple behaviors from one task
With Tatsuya Harada, Osa coauthored an ICML paper on discovering multiple solutions from a single task in offline reinforcement learning. He directs RIKEN’s Robot Learning Team.
Sources and credits
Photo: www.riken.jp; Takayuki Osa profile photograph; creator not stated. License: Not stated by source. Source: riken.jp. Cropped to a square and resized. Ask us to remove or replace this photo.