IEICE Technical Committee Submission System
Conference Paper's Information
Online Proceedings
[Sign in]
... (for ESS/CS/ES/ISS)
Tech. Rep. Archives
... (for ES/CS)
 Go Top Page Go Previous   [Japanese] / [English] 

Paper Abstract and Keywords
Presentation 2019-03-15 14:55
On the Influence of Momentum term in quasi-Newton method
Shahrzad Mahboubi (SIT), Indrapriyadarsini s (Shizuoka Univ.), Hiroshi Ninomiya (SIT), Hideki Asai (Shizuoka Univ.)
Abstract (in Japanese) (See Japanese page) 
(in English) The Nesterov's Accelerated quasi-Newton (NAQ) method was derived from the quadratic approximation of the error function incorporating the momentum term in the quasi-Newton method. This method was effective for training in neural networks. However, the gradient calculation of NAQ was conducted two times within one iteration. This increased the computation time. In this research we propose a new quasi-Newton based training method that approximates the Nesterov's accelerated gradient used in NAQ such that the gradient is computed only once per iteration. The performance of the proposed method is evaluated on benchmark data-sets and microwave circuit modeling data. The results show significant acceleration in computation time compared to conventional training algorithms.
Keyword (in Japanese) (See Japanese page) 
(in English) Neural network / training algorithm / quasi-Newton method / Nesterov's accelerated quasi-Newton method / momentum term / / /  
Reference Info. IEICE Tech. Rep., vol. 118, no. 498, NLP2018-137, pp. 69-74, March 2019.
Paper # NLP2018-137 
Date of Issue 2019-03-07 (NLP) 
ISSN Print edition: ISSN 0913-5685  Online edition: ISSN 2432-6380

Conference Information
Committee NLP MSS  
Conference Date 2019-03-14 - 2019-03-15 
Place (in Japanese) (See Japanese page) 
Place (in English) Bunkyo Camp., Univ. of Fukui 
Topics (in Japanese) (See Japanese page) 
Topics (in English) SICE-DES, IEICE-MSS, IEICE-NLP, Work In Progress, and etc. 
Paper Information
Registration To NLP 
Conference Code 2019-03-NLP-MSS 
Language Japanese 
Title (in Japanese) (See Japanese page) 
Sub Title (in Japanese) (See Japanese page) 
Title (in English) On the Influence of Momentum term in quasi-Newton method 
Sub Title (in English)  
Keyword(1) Neural network  
Keyword(2) training algorithm  
Keyword(3) quasi-Newton method  
Keyword(4) Nesterov's accelerated quasi-Newton method  
Keyword(5) momentum term  
1st Author's Name Shahrzad Mahboubi  
1st Author's Affiliation Shonan Institute of Technology (SIT)
2nd Author's Name Indrapriyadarsini s  
2nd Author's Affiliation Shizuoka University (Shizuoka Univ.)
3rd Author's Name Hiroshi Ninomiya  
3rd Author's Affiliation Shonan Institute of Technology (SIT)
4th Author's Name Hideki Asai  
4th Author's Affiliation Shizuoka University (Shizuoka Univ.)
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 ()
Date Time 2019-03-15 14:55:00 
Presentation Time 25 
Registration for NLP 
Paper # IEICE-NLP2018-137 
Volume (vol) IEICE-118 
Number (no) no.498 
Page pp.69-74 
#Pages IEICE-6 
Date of Issue IEICE-NLP-2019-03-07 

[Return to Top Page]

[Return to IEICE Web Page]

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