Paper Abstract and Keywords |
Presentation |
2018-03-13 10:50
Statistical mechanics of coherent Ising machine
-- The analysis of Hopfield model with discrete synapses -- Shunki Nakagawa, Toru Aonishi (Tokyo Tech.), Kazushi Mimura (Hiroshima City Univ.), Masato Okada (Univ. Tokyo), Yoshihisa Yamamoto (ImPACT/ Stanford Univ.) NC2017-70 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
The coherent Ising machine (CIM) is being developed as one of the Ising computing architecture for solving large-scale combinatorial optimization problems at high speed. However, because the CIM is a non-equilibrium open dissipative system, it is not possible to directly apply an equilibrium statistical mechanics method to the CIM. Furthermore, in the actual equipment, synapses can take only discrete values of -1, 0, +1. It is not clear how this quantization affects the performance of the system. Here, we focus on the CIM-implemented Hopfield model with discrete synapses. Using our statistical mechanics method, we drive a macroscopic equation to elucidate the behavior of critical memory capacity under various values of pump rate. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
coherent Ising machine / statistical mechanics / Hopfield model / macroscopic equation / critical memory capacity / / / |
Reference Info. |
IEICE Tech. Rep., vol. 117, no. 508, NC2017-70, pp. 15-20, March 2018. |
Paper # |
NC2017-70 |
Date of Issue |
2018-03-06 (NC) |
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 |
NC2017-70 |
|