PVEs: Position-Velocity Encoders for Unsupervised Learning of Structured State Representations
Abstract
We propose position-velocity encoders (PVEs) which learn---without supervision---to encode images to positions and velocities of task-relevant objects. PVEs encode a single image into a low-dimensional position state and compute the velocity state from finite differences in position. In contrast to autoencoders, position-velocity encoders are not trained by image reconstruction, but by making the position-velocity representation consistent with priors about interacting with the physical world. We applied PVEs to several simulated control tasks from pixels and achieved promising preliminary results.
Keywords
Cite
@article{arxiv.1705.09805,
title = {PVEs: Position-Velocity Encoders for Unsupervised Learning of Structured State Representations},
author = {Rico Jonschkowski and Roland Hafner and Jonathan Scholz and Martin Riedmiller},
journal= {arXiv preprint arXiv:1705.09805},
year = {2017}
}
Comments
Accepted at Robotics: Science and Systems (RSS 2017) Workshop -- New Frontiers for Deep Learning in Robotics http://juxi.net/workshop/deep-learning-rss-2017/