Cover Image

Emergence of discrete and abstract state representation through reinforcement learning in a continuous input task

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

Saved in:
Bibliographic Details
Main Authors: Sawatsubashi, Yoshito, Mohamad Faizal, Samsudin, Shibata, Katsunari
Other Authors: bashis8@yahoo.co.jp
Format: Article
Language:English
Published: Springer-Verlag 2014
Subjects:
Action planning
Concept formation
Continuous input
Hidden neurons
Online Access:http://dspace.unimap.edu.my:80/dspace/handle/123456789/35395
Tags: Add Tag
No Tags, Be the first to tag this record!
id my.unimap-35395
record_format dspace
spelling my.unimap-353952014-06-11T13:45:49Z Emergence of discrete and abstract state representation through reinforcement learning in a continuous input task Sawatsubashi, Yoshito Mohamad Faizal, Samsudin Shibata, Katsunari bashis8@yahoo.co.jp ballack83@hotmail.co.jp faizalsamsudin@unimap.edu.my Action planning Concept formation Continuous input Hidden neurons Link to publisher's homepage at http://link.springer.com/ "Concept" is a kind of discrete and abstract state representation, and is considered useful for efficient action planning. However, it is supposed to emerge in our brain as a parallel processing and learning system through learning based on a variety of experiences, and so it is difficult to be developed by hand-coding. In this paper, as a previous step of the "concept formation", it is investigated whether the discrete and abstract state representation is formed or not through learning in a task with multi-step state transitions using Actor-Q learning method and a recurrent neural network. After learning, an agent repeated a sequence two times, in which it pushed a button to open a door and moved to the next room, and finally arrived at the third room to get a reward. In two hidden neurons, discrete and abstract state representation not depending on the door opening pattern was observed. The result of another learning with two recurrent neural networks that are for Q-values and for Actors suggested that the state representation emerged to generate appropriate Q-values. 2014-06-11T13:45:49Z 2014-06-11T13:45:49Z 2013 Article Advances in Intelligent Systems and Computing, vol. 208, 2013, pages 13-21 978-364237373-2 2194-5357 http://link.springer.com/chapter/10.1007%2F978-3-642-37374-9_2 http://dspace.unimap.edu.my:80/dspace/handle/123456789/35395 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 Action planning
Concept formation
Continuous input
Hidden neurons
spellingShingle Action planning
Concept formation
Continuous input
Hidden neurons
Sawatsubashi, Yoshito
Mohamad Faizal, Samsudin
Shibata, Katsunari
Emergence of discrete and abstract state representation through reinforcement learning in a continuous input task
description Link to publisher's homepage at http://link.springer.com/
author2 bashis8@yahoo.co.jp
author_facet bashis8@yahoo.co.jp
Sawatsubashi, Yoshito
Mohamad Faizal, Samsudin
Shibata, Katsunari
format Article
author Sawatsubashi, Yoshito
Mohamad Faizal, Samsudin
Shibata, Katsunari
author_sort Sawatsubashi, Yoshito
title Emergence of discrete and abstract state representation through reinforcement learning in a continuous input task
title_short Emergence of discrete and abstract state representation through reinforcement learning in a continuous input task
title_full Emergence of discrete and abstract state representation through reinforcement learning in a continuous input task
title_fullStr Emergence of discrete and abstract state representation through reinforcement learning in a continuous input task
title_full_unstemmed Emergence of discrete and abstract state representation through reinforcement learning in a continuous input task
title_sort emergence of discrete and abstract state representation through reinforcement learning in a continuous input task
publisher Springer-Verlag
publishDate 2014
url http://dspace.unimap.edu.my:80/dspace/handle/123456789/35395
_version_ 1643797782117482496
score 13.211869

Similar Items