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 2022-09-08 10:10
Proposal and evaluation of Combined Posit MAC unit (CPMAC) for both DNN inference and training
Yuta Masuda, Yasuhiro Nakahara, Masato Kiyama, Masahiro Iida (Kumamoto Univ.) RECONF2022-34
Abstract (in Japanese) (See Japanese page) 
(in English) Recently, there has been a lot of research on DNN hardware accelerators for the edge that use Posit as a number representation. Although the Posit contributes highly accurate inference and training using a fewer bit-width than floating point numbers, there is a difference in the required bits accuracy between inference and training, requiring other arithmetic units for each. However, it is difficult to implement both arithmetic units on edge devices with limited resources. In this article, we propose a Combined Posit MAC unit (CPMAC) for inference and training which can combine a lower precision Posit MAC
unit with the plural. As a result, we achieved an area reduction of more than 20% at the maximum when the exponent of Posit is large, and demonstrated the usefulness of CPMAC in applications that require a wide dynamic range.
Keyword (in Japanese) (See Japanese page) 
(in English) DeeoLearning / Convolutional Neural Network / Posit / MAC unit / / / /  
Reference Info. IEICE Tech. Rep., vol. 122, no. 174, RECONF2022-34, pp. 29-34, Sept. 2022.
Paper # RECONF2022-34 
Date of Issue 2022-08-31 (RECONF) 
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)
Notes on Review This article is a technical report without peer review, and its polished version will be published elsewhere.
Download PDF RECONF2022-34

Conference Information
Committee RECONF  
Conference Date 2022-09-07 - 2022-09-08 
Place (in Japanese) (See Japanese page) 
Place (in English) emCAMPUS STUDIO 
Topics (in Japanese) (See Japanese page) 
Topics (in English) Reconfigurable system, etc. 
Paper Information
Registration To RECONF 
Conference Code 2022-09-RECONF 
Language Japanese 
Title (in Japanese) (See Japanese page) 
Sub Title (in Japanese) (See Japanese page) 
Title (in English) Proposal and evaluation of Combined Posit MAC unit (CPMAC) for both DNN inference and training 
Sub Title (in English)  
Keyword(1) DeeoLearning  
Keyword(2) Convolutional Neural Network  
Keyword(3) Posit  
Keyword(4) MAC unit  
Keyword(5)  
Keyword(6)  
Keyword(7)  
Keyword(8)  
1st Author's Name Yuta Masuda  
1st Author's Affiliation Kumamoto University (Kumamoto Univ.)
2nd Author's Name Yasuhiro Nakahara  
2nd Author's Affiliation Kumamoto University (Kumamoto Univ.)
3rd Author's Name Masato Kiyama  
3rd Author's Affiliation Kumamoto University (Kumamoto Univ.)
4th Author's Name Masahiro Iida  
4th Author's Affiliation Kumamoto University (Kumamoto 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 ()
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 2022-09-08 10:10:00 
Presentation Time 25 minutes 
Registration for RECONF 
Paper # RECONF2022-34 
Volume (vol) vol.122 
Number (no) no.174 
Page pp.29-34 
#Pages 6 
Date of Issue 2022-08-31 (RECONF) 

[Return to Top Page]

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