Cover Image

Emergence of multi-step discrete state transition through reinforcement learning with a recurrent neural network

Link to publisher's homepage at http://link.springer.com/

Saved in:
Bibliographic Details
Main Authors: Mohamad Faizal, Samsudin, Sawatsubashi, Yoshito, Katsunari, Shibata
Other Authors: faizalsamsudin@unimap.edu.my
Format: Article
Language:English
Published: Springer-Verlag 2014
Subjects:
Multi-step discrete state transition
Recurrent neural network
Reinforcement learning
Online Access:http://dspace.unimap.edu.my/xmlui/handle/123456789/30948
Tags: Add Tag
No Tags, Be the first to tag this record!
id my.unimap-30948
record_format dspace
spelling my.unimap-309482014-01-10T08:55:22Z Emergence of multi-step discrete state transition through reinforcement learning with a recurrent neural network Mohamad Faizal, Samsudin Sawatsubashi, Yoshito Katsunari, Shibata faizalsamsudin@unimap.edu.my bashis8@yahoo.co.jp shibata@oita-u.ac.jp Multi-step discrete state transition Recurrent neural network Reinforcement learning Link to publisher's homepage at http://link.springer.com/ For developing a robot that learns long and complicated action sequences in the real-world, autonomous learning of multi-step discrete state transition is significant. To realize the multi-step discrete state transition in a neural network is generally thought to be difficult because of basically the needs to hold the state while performing the transition between the states when needed. In this paper, only through the reinforcement learning using rewards and punishments in a simple learning system consisting of a recurrent neural network (RNN), it is shown that a multi-step discrete state transition emerged through learning in a continuous state-action space. It is shown that in a two-switch task, two states transition represented by two types of hidden nodes emerged through the learning. In addition, it is shown that the contribution of the dynamics by the interaction between the RNN and the environment based on the discrete state transitions leads to repetition of the interesting behavior when no reward is given at the goal. 2014-01-02T08:46:59Z 2014-01-02T08:46:59Z 2012 Article Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 7664 LNCS(Part2), 2012, pages 583-590 978-364234480-0 0302-9743 http://link.springer.com/chapter/10.1007%2F978-3-642-34481-7_71 http://hdl.handle.net/123456789/30948 en Springer-Verlag
institution Universiti Malaysia Perlis
building UniMAP Library
collection Institutional Repository
continent Asia
country Malaysia
content_provider Universiti Malaysia Perlis
content_source UniMAP Library Digital Repository
url_provider http://dspace.unimap.edu.my/
language English
topic Multi-step discrete state transition
Recurrent neural network
Reinforcement learning
spellingShingle Multi-step discrete state transition
Recurrent neural network
Reinforcement learning
Mohamad Faizal, Samsudin
Sawatsubashi, Yoshito
Katsunari, Shibata
Emergence of multi-step discrete state transition through reinforcement learning with a recurrent neural network
description Link to publisher's homepage at http://link.springer.com/
author2 faizalsamsudin@unimap.edu.my
author_facet faizalsamsudin@unimap.edu.my
Mohamad Faizal, Samsudin
Sawatsubashi, Yoshito
Katsunari, Shibata
format Article
author Mohamad Faizal, Samsudin
Sawatsubashi, Yoshito
Katsunari, Shibata
author_sort Mohamad Faizal, Samsudin
title Emergence of multi-step discrete state transition through reinforcement learning with a recurrent neural network
title_short Emergence of multi-step discrete state transition through reinforcement learning with a recurrent neural network
title_full Emergence of multi-step discrete state transition through reinforcement learning with a recurrent neural network
title_fullStr Emergence of multi-step discrete state transition through reinforcement learning with a recurrent neural network
title_full_unstemmed Emergence of multi-step discrete state transition through reinforcement learning with a recurrent neural network
title_sort emergence of multi-step discrete state transition through reinforcement learning with a recurrent neural network
publisher Springer-Verlag
publishDate 2014
url http://dspace.unimap.edu.my/xmlui/handle/123456789/30948
_version_ 1643796421056397312
score 13.214268

Similar Items