State representation learning with recurrent capsule networks
Machine Learning
2019-02-25 v4 Neural and Evolutionary Computing
Abstract
Unsupervised learning of compact and relevant state representations has been proved very useful at solving complex reinforcement learning tasks. In this paper, we propose a recurrent capsule network that learns such representations by trying to predict the future observations in an agent's trajectory.
Cite
@article{arxiv.1812.11202,
title = {State representation learning with recurrent capsule networks},
author = {Louis Annabi and Michael Garcia Ortiz},
journal= {arXiv preprint arXiv:1812.11202},
year = {2019}
}
Comments
4 pages, 4 figures, NIPS Workshop on Modeling the Physical World: Perception, Learning, and Control