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 2017-03-07 10:30
Doubly Accelerated Stochastic Variance Reduced Gradient Method for Regularized Empirical Risk Minimization
Tomoya Murata, Taiji Suzuki (Tokyo Tech) IBISML2016-106
Abstract (in Japanese) (See Japanese page) 
(in English) We develop a new stochastic gradient method for solving convex regularized empirical risk minimization problem in mini-batch settings. The core of our proposed method is incorporation of our new technique ``double acceleration'' and variance reduction technique. We theoretically analyze our proposed method, and show that our method much improves the mini-batch efficiency of previous accelerated stochastic methods, and essentially only needs size $sqrt{n}$ mini-batches for achieving the optimal iteration complexities for both non-strongly and strongly convex objectives, where $n$ is the training set size. Furthermore, we show that even in non-mini-batch settings, our method still improves the best known convergence rate for non-strongly convex objectives, and achieves the one for strongly convex objectives.
Keyword (in Japanese) (See Japanese page) 
(in English) convex optimizaition / empirical risk minimization / stochastic optimization / variance reduction / double acceleration / mini-batch method / /  
Reference Info. IEICE Tech. Rep., vol. 116, no. 500, IBISML2016-106, pp. 49-56, March 2017.
Paper # IBISML2016-106 
Date of Issue 2017-02-27 (IBISML) 
ISSN Print edition: ISSN 0913-5685    Online edition: ISSN 2432-6380
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 IBISML2016-106

Conference Information
Committee IBISML  
Conference Date 2017-03-06 - 2017-03-07 
Place (in Japanese) (See Japanese page) 
Place (in English) Tokyo Institute of Technology 
Topics (in Japanese) (See Japanese page) 
Topics (in English) Statistical Mathematics, Machine Learning, Data Mining, etc. 
Paper Information
Registration To IBISML 
Conference Code 2017-03-IBISML 
Language English (Japanese title is available) 
Title (in Japanese) (See Japanese page) 
Sub Title (in Japanese) (See Japanese page) 
Title (in English) Doubly Accelerated Stochastic Variance Reduced Gradient Method for Regularized Empirical Risk Minimization 
Sub Title (in English)  
Keyword(1) convex optimizaition  
Keyword(2) empirical risk minimization  
Keyword(3) stochastic optimization  
Keyword(4) variance reduction  
Keyword(5) double acceleration  
Keyword(6) mini-batch method  
1st Author's Name Tomoya Murata  
1st Author's Affiliation Tokyo Institute of Technology (Tokyo Tech)
2nd Author's Name Taiji Suzuki  
2nd Author's Affiliation Tokyo Institute of Technology (Tokyo Tech)
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 2017-03-07 10:30:00 
Presentation Time 30 minutes 
Registration for IBISML 
Paper # IBISML2016-106 
Volume (vol) vol.116 
Number (no) no.500 
Page pp.49-56 
Date of Issue 2017-02-27 (IBISML) 

[Return to Top Page]

[Return to IEICE Web Page]

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