Paper Abstract and Keywords |
Presentation |
2016-11-17 14:00
[Poster Presentation]
Estimating Proton Conductivity in Crystals by using Guassian Process and Dynamic Programming Kenta Kanamori (NITech), Kazuaki Toyoura (Kyoto Univ.), Shinichi Nakajima (TU Berlin), Atsuto Seko (Kyoto Univ.), Masayuki Karasuyama (NITech), Akihide Kuwabara (JFCC), Junya Honda (Tokyo Univ.), Kazuki Shitara (JFCC), Motoki Shiga (Gifu Univ.), Ichiro Takeuchi (NITech) IBISML2016-73 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
In material science, $proton conductivity$ is very important property for designing new battery and it is defined as maximum $potential energy$ in the proton moving path over $potential energy surface$ (PES).
Many material scientists usually calculate proton conductivity by using simulation on PES.
However, to know the whole exact PES is very time consuming.
In this paper, we propose a method for proton coducitivity estimation which combines the probabilistic modeling of PES by Gaussian process with dynamic programming.
We also propose a selective sampling for improving estimation accuracy.
Our experimental result with three type of proton conductor shows the proposed approach brings significant reduction of computational cost. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
Proton Conductivity / Gaussian Process / Dynamic Programming / Bottle-neck Search / / / / |
Reference Info. |
IEICE Tech. Rep., vol. 116, no. 300, IBISML2016-73, pp. 191-198, Nov. 2016. |
Paper # |
IBISML2016-73 |
Date of Issue |
2016-11-09 (IBISML) |
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 |
IBISML2016-73 |
|