Information: Join today and make your research activities more affordable! Technical workshop participation fees and annual registration fees are available at member rates.
Notice: [Important] Announcement of Changes to Registration Fee Payment and Manuscript Upload Procedures for IEICE Technical Meetings
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 2022-01-21 16:00
On the Study of Second-Order Training Algorithm using Matrix Diagonalization based on Hutchinson estimation
Ryo Yamatomi, Shahrzad Mahboubi, Hiroshi Ninomiya (Shonan Inst. Tec.) NLP2021-89 MICT2021-64 MBE2021-50
Abstract (in Japanese) (See Japanese page) 
(in English) In this study, we propose a new training algorithm based on the second-order approximated gradient method, which aims to reduce the computational cost of the Newton method. In neural network (NN) training, the first-order approximate gradient methods are commonly used because of its low computational cost. However, when applied to highly nonlinear problems, first-order methods still converge too slowly, and optimization error cannot effectively reduced within finite time despite its advantages. On the other hand, the training algorithm based on Newton method which is considered effective for this problem, is computationally expensive because it uses Hessian matrices, making it difficult to train for large-scale nonlinear problems. In this study, we focus on reducing the computational cost of the Newton method and propose a new learning algorithm as Hutchinson diagonalized Newton method (HdN), which is realized by the approximated diagonal Hessian matrix using the diagonal matrix based on Hutchinson estimator. We apply the proposed method to the training of NN, and the performance of the proposed method is demonstrated through computer simulations.
Keyword (in Japanese) (See Japanese page) 
(in English) Neural network / training algorithm / Newton method / hessian approximation scheme / Hutchinson estimator / / /  
Reference Info. IEICE Tech. Rep., vol. 121, no. 335, NLP2021-89, pp. 67-70, Jan. 2022.
Paper # NLP2021-89 
Date of Issue 2022-01-14 (NLP, MICT, MBE) 
ISSN 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 NLP2021-89 MICT2021-64 MBE2021-50

Conference Information
Committee NLP MICT MBE NC  
Conference Date 2022-01-21 - 2022-01-23 
Place (in Japanese) (See Japanese page) 
Place (in English) Online 
Topics (in Japanese) (See Japanese page) 
Topics (in English)  
Paper Information
Registration To NLP 
Conference Code 2022-01-NLP-MICT-MBE-NC 
Language Japanese 
Title (in Japanese) (See Japanese page) 
Sub Title (in Japanese) (See Japanese page) 
Title (in English) On the Study of Second-Order Training Algorithm using Matrix Diagonalization based on Hutchinson estimation 
Sub Title (in English)  
Keyword(1) Neural network  
Keyword(2) training algorithm  
Keyword(3) Newton method  
Keyword(4) hessian approximation scheme  
Keyword(5) Hutchinson estimator  
Keyword(6)  
Keyword(7)  
Keyword(8)  
1st Author's Name Ryo Yamatomi  
1st Author's Affiliation Shonan Institute of Technology (Shonan Inst. Tec.)
2nd Author's Name Shahrzad Mahboubi  
2nd Author's Affiliation Shonan Institute of Technology (Shonan Inst. Tec.)
3rd Author's Name Hiroshi Ninomiya  
3rd Author's Affiliation Shonan Institute of Technology (Shonan Inst. Tec.)
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 ()
21st Author's Name  
21st Author's Affiliation ()
22nd Author's Name  
22nd Author's Affiliation ()
23rd Author's Name  
23rd Author's Affiliation ()
24th Author's Name  
24th Author's Affiliation ()
25th Author's Name  
25th Author's Affiliation ()
26th Author's Name / /
26th Author's Affiliation ()
()
27th Author's Name / /
27th Author's Affiliation ()
()
28th Author's Name / /
28th Author's Affiliation ()
()
29th Author's Name / /
29th Author's Affiliation ()
()
30th Author's Name / /
30th Author's Affiliation ()
()
31st Author's Name / /
31st Author's Affiliation ()
()
32nd Author's Name / /
32nd Author's Affiliation ()
()
33rd Author's Name / /
33rd Author's Affiliation ()
()
34th Author's Name / /
34th Author's Affiliation ()
()
35th Author's Name / /
35th Author's Affiliation ()
()
36th Author's Name / /
36th Author's Affiliation ()
()
Speaker Author-1 
Date Time 2022-01-21 16:00:00 
Presentation Time 25 minutes 
Registration for NLP 
Paper # NLP2021-89, MICT2021-64, MBE2021-50 
Volume (vol) vol.121 
Number (no) no.335(NLP), no.336(MICT), no.337(MBE) 
Page pp.67-70 
#Pages 4 
Date of Issue 2022-01-14 (NLP, MICT, MBE) 

[Return to Top Page]

[Return to IEICE Web Page]
The Institute of Electronics, Information and Communication Engineers (IEICE), Japan