Paper Abstract and Keywords |
Presentation |
2020-12-18 14:25
Zero-shot generative model considering attribute uncertainty Yuta Sakai (Waseda Univ.), Kenta Mikawa (SIT), Masayuki Goto (Waseda Univ.) PRMU2020-59 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
Classification problems in machine learning remain an important research topic. In general, classification estimates unknown labels of test data by using a training data set that consists of pairs of data features and labels representing categories. However, in general classification problems, it is not possible to estimate unkown categories that are not included in the training data set. In this research, we focus on zero-shot learning, which is a method that makes it possible to estimate categories that are not in the learning data by utilizing information (auxiliary information) that is common to the data. Among them, we pay particular attention to attribute-based zero-shot learning that utilizes attribute information as auxiliary information.
The Direct Attribute Prediction (DAP) model, which is one of the conventional methods for performing attribute-based zero-shot learning, is a discriminative model that expresses the relationship between features and categories by utilizing attributes. However, in the DAP model, estimation results other than the important attributes for category estimation may reduce the estimation accuracy of the category. Moreover, since it is a discriminative model, it may be overfitted when the number of training data is small.
Therefore, in this study, we propose an attribute-based zero-shot generation model that takes into account the uncertainty of the estimation result of the attribute information. Finally, we verify the effectiveness of the proposed method comparing the conventional method. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
Zero-shot Learning / Attribute information / generative model / classification / / / / |
Reference Info. |
IEICE Tech. Rep., vol. 120, no. 300, PRMU2020-59, pp. 122-127, Dec. 2020. |
Paper # |
PRMU2020-59 |
Date of Issue |
2020-12-10 (PRMU) |
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 |
PRMU2020-59 |
Conference Information |
Committee |
PRMU |
Conference Date |
2020-12-17 - 2020-12-18 |
Place (in Japanese) |
(See Japanese page) |
Place (in English) |
Online |
Topics (in Japanese) |
(See Japanese page) |
Topics (in English) |
Transfer learning and few shot learning |
Paper Information |
Registration To |
PRMU |
Conference Code |
2020-12-PRMU |
Language |
Japanese |
Title (in Japanese) |
(See Japanese page) |
Sub Title (in Japanese) |
(See Japanese page) |
Title (in English) |
Zero-shot generative model considering attribute uncertainty |
Sub Title (in English) |
|
Keyword(1) |
Zero-shot Learning |
Keyword(2) |
Attribute information |
Keyword(3) |
generative model |
Keyword(4) |
classification |
Keyword(5) |
|
Keyword(6) |
|
Keyword(7) |
|
Keyword(8) |
|
1st Author's Name |
Yuta Sakai |
1st Author's Affiliation |
Waseda University (Waseda Univ.) |
2nd Author's Name |
Kenta Mikawa |
2nd Author's Affiliation |
Shonan Institute of Technology (SIT) |
3rd Author's Name |
Masayuki Goto |
3rd Author's Affiliation |
Waseda University (Waseda Univ.) |
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 |
() |
Speaker |
Author-1 |
Date Time |
2020-12-18 14:25:00 |
Presentation Time |
15 minutes |
Registration for |
PRMU |
Paper # |
PRMU2020-59 |
Volume (vol) |
vol.120 |
Number (no) |
no.300 |
Page |
pp.122-127 |
#Pages |
6 |
Date of Issue |
2020-12-10 (PRMU) |
|