Please use this identifier to cite or link to this item: http://dspace.unimap.edu.my:80/xmlui/handle/123456789/30948
Full metadata record
DC FieldValueLanguage
dc.contributor.authorMohamad Faizal, Samsudin-
dc.contributor.authorSawatsubashi, Yoshito-
dc.contributor.authorKatsunari, Shibata-
dc.date.accessioned2014-01-02T08:46:59Z-
dc.date.available2014-01-02T08:46:59Z-
dc.date.issued2012-
dc.identifier.citationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 7664 LNCS(Part2), 2012, pages 583-590en_US
dc.identifier.isbn978-364234480-0-
dc.identifier.issn0302-9743-
dc.identifier.urihttp://link.springer.com/chapter/10.1007%2F978-3-642-34481-7_71-
dc.identifier.urihttp://dspace.unimap.edu.my/123456789/30948-
dc.descriptionLink to publisher's homepage at http://link.springer.com/en_US
dc.description.abstractFor 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.en_US
dc.language.isoenen_US
dc.publisherSpringer-Verlagen_US
dc.subjectMulti-step discrete state transitionen_US
dc.subjectRecurrent neural networken_US
dc.subjectReinforcement learningen_US
dc.titleEmergence of multi-step discrete state transition through reinforcement learning with a recurrent neural networken_US
dc.typeArticleen_US
dc.contributor.urlfaizalsamsudin@unimap.edu.myen_US
dc.contributor.urlbashis8@yahoo.co.jpen_US
dc.contributor.urlshibata@oita-u.ac.jpen_US
Appears in Collections:School of Mechatronic Engineering (Articles)



Items in UniMAP Library Digital Repository are protected by copyright, with all rights reserved, unless otherwise indicated.