A0C: Alpha Zero in Continuous Action Space
Machine Learning
2018-05-25 v1 Artificial Intelligence
Machine Learning
Robotics
Systems and Control
Abstract
A core novelty of Alpha Zero is the interleaving of tree search and deep learning, which has proven very successful in board games like Chess, Shogi and Go. These games have a discrete action space. However, many real-world reinforcement learning domains have continuous action spaces, for example in robotic control, navigation and self-driving cars. This paper presents the necessary theoretical extensions of Alpha Zero to deal with continuous action space. We also provide some preliminary experiments on the Pendulum swing-up task, empirically showing the feasibility of our approach. Thereby, this work provides a first step towards the application of iterated search and learning in domains with a continuous action space.
Cite
@article{arxiv.1805.09613,
title = {A0C: Alpha Zero in Continuous Action Space},
author = {Thomas M. Moerland and Joost Broekens and Aske Plaat and Catholijn M. Jonker},
journal= {arXiv preprint arXiv:1805.09613},
year = {2018}
}