IEICE Technical Committee Submission System
Conference Paper's Information
Online Proceedings
[Sign in]
Tech. Rep. Archives
 Go Top Page Go Previous   [Japanese] / [English] 

Paper Abstract and Keywords
Presentation 2012-03-12 15:05
Model Selection of Indirect Value Function Estimation
Masahiro Kohjima (Tokyo Tech) IBISML2011-93
Abstract (in Japanese) (See Japanese page) 
(in English) Reinforcement learning is a method to obtain a policy which maximizes expected return and is applied to wide range of research fields such as robotics, recommendation system, and natural language processing. Estimation of value function is of importance to realize reinforcement learning. In this study, the author investigates a method to compute value function indirectly through an estimation of transition probability. The asymptotic behavior of mean squared error of true and estimated value function is theoretically clarified in the case where transition probability parameter has a structure. The author also proposes a model selection criterion based on this analysis and verifies its effectiveness by numerical experiment.
Keyword (in Japanese) (See Japanese page) 
(in English) reinforcement learning / markov decision processes / transition probability estimation / model selection / / / /  
Reference Info. IEICE Tech. Rep., vol. 111, no. 480, IBISML2011-93, pp. 43-48, March 2012.
Paper # IBISML2011-93 
Date of Issue 2012-03-05 (IBISML) 
ISSN Print edition: ISSN 0913-5685    Online edition: ISSN 2432-6380
Copyright
and
reproduction
All rights are reserved and no part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Notwithstanding, instructors are permitted to photocopy isolated articles for noncommercial classroom use without fee. (License No.: 10GA0019/12GB0052/13GB0056/17GB0034/18GB0034)
Download PDF IBISML2011-93

Conference Information
Committee IBISML  
Conference Date 2012-03-12 - 2012-03-13 
Place (in Japanese) (See Japanese page) 
Place (in English) The Institute of Statistical Mathematics 
Topics (in Japanese) (See Japanese page) 
Topics (in English) Mathematical Statistics, Databases, etc. 
Paper Information
Registration To IBISML 
Conference Code 2012-03-IBISML 
Language Japanese 
Title (in Japanese) (See Japanese page) 
Sub Title (in Japanese) (See Japanese page) 
Title (in English) Model Selection of Indirect Value Function Estimation 
Sub Title (in English)  
Keyword(1) reinforcement learning  
Keyword(2) markov decision processes  
Keyword(3) transition probability estimation  
Keyword(4) model selection  
Keyword(5)  
Keyword(6)  
Keyword(7)  
Keyword(8)  
1st Author's Name Masahiro Kohjima  
1st Author's Affiliation Tokyo Institute Of Technology (Tokyo Tech)
2nd Author's Name  
2nd Author's Affiliation ()
3rd Author's Name  
3rd Author's Affiliation ()
4th Author's Name  
4th Author's Affiliation ()
5th Author's Name  
5th Author's Affiliation ()
6th Author's Name  
6th Author's Affiliation ()
7th Author's Name  
7th Author's Affiliation ()
8th Author's Name  
8th Author's Affiliation ()
9th Author's Name  
9th Author's Affiliation ()
10th Author's Name  
10th Author's Affiliation ()
11th Author's Name  
11th Author's Affiliation ()
12th Author's Name  
12th Author's Affiliation ()
13th Author's Name  
13th Author's Affiliation ()
14th Author's Name  
14th Author's Affiliation ()
15th Author's Name  
15th Author's Affiliation ()
16th Author's Name  
16th Author's Affiliation ()
17th Author's Name  
17th Author's Affiliation ()
18th Author's Name  
18th Author's Affiliation ()
19th Author's Name  
19th Author's Affiliation ()
20th Author's Name  
20th Author's Affiliation ()
Speaker Author-1 
Date Time 2012-03-12 15:05:00 
Presentation Time 25 minutes 
Registration for IBISML 
Paper # IBISML2011-93 
Volume (vol) vol.111 
Number (no) no.480 
Page pp.43-48 
#Pages
Date of Issue 2012-03-05 (IBISML) 


[Return to Top Page]

[Return to IEICE Web Page]


The Institute of Electronics, Information and Communication Engineers (IEICE), Japan