S
Sethu Vijayakumar
Researcher at University of Edinburgh
Publications - 361
Citations - 10209
Sethu Vijayakumar is an academic researcher from University of Edinburgh. The author has contributed to research in topics: Robot & Humanoid robot. The author has an hindex of 42, co-authored 338 publications receiving 8921 citations. Previous affiliations of Sethu Vijayakumar include Laboratory for Analysis and Architecture of Systems & University of Oxford.
Papers
More filters
Proceedings Article
Natural Actor-Critic
TL;DR: This paper investigates a novel model-free reinforcement learning architecture, the Natural Actor-Critic, where the actor updates are based on stochastic policy gradients employing Amari's natural gradient approach, while the critic obtains both the natural policy gradient and additional parameters of a value function simultaneously by linear regression.
Journal ArticleDOI
Incremental Online Learning in High Dimensions
TL;DR: Locally weighted projection regression is the first truly incremental spatially localized learning method that can successfully and efficiently operate in very high-dimensional spaces.
Proceedings Article
Reinforcement Learning for Humanoid Robotics
TL;DR: This paper discusses different approaches of reinforcement learning in terms of their applicability in humanoid robotics, and demonstrates that ‘vanilla’ policy gradient methods can be significantly improved using the natural policy gradient instead of the regular policy gradient.
Locally Weighted Projection Regression : An O(n) Algorithm for Incremental Real Time Learning in High Dimensional Space
Sethu Vijayakumar,Stefan Schaal +1 more
TL;DR: This paper evaluates different methods of projection regression and derives a nonlinear function approximator based on them, which is the first truly incremental spatially localized learning method to combine all these properties.
Book ChapterDOI
Natural actor-critic
TL;DR: The Natural Actor-Critic as mentioned in this paper is a model-free reinforcement learning architecture, where actor updates are based on stochastic policy gradients employing Amari's natural gradient approach, while the critic obtains both the natural policy gradient and additional parameters of a value function simultaneously by linear regression.