1207631
doi
10.5281/zenodo.1207631
oai:zenodo.org:1207631
Schmidhuber, Jürgen
NNAISENSE
World Models
Ha, David
Google Brain
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
<p>We explore building generative neural network models of popular reinforcement learning environments. Our <em>world model</em> can be trained quickly in an unsupervised manner to learn a compressed spatial and temporal representation of the environment. By using features extracted from the world model as inputs to an agent, we can train a very compact and simple policy that can solve the required task. We can even train our agent entirely inside of its own hallucinated dream generated by its world model, and transfer this policy back into the actual environment.</p>
<p>An interactive version of this article is available at <a href="https://worldmodels.github.io">worldmodels.github.io</a>.</p>
Interactive version of the article at https://worldmodels.github.io
Zenodo
2018-03-28
info:eu-repo/semantics/article
1207048
1.1
1579540888.086057
2980722
md5:781c6e1b5ad16f89739c48371148c121
https://zenodo.org/records/1207631/files/WorldModelsArxiv.pdf
public
10.5281/zenodo.1207048
isVersionOf
doi