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 2026-07-21 16:40
Augmenting Minority Emotion Labels in Japanese Emotion Classification through Translation of English Emotion-Labeled Data: Effectiveness and Limitations in WRIME
Yuya Wake, Sho Tsugawa (Univ. of Tsukuba) CQ2026-33
Abstract (in Japanese) (See Japanese page) 
(in English) WRIME is an emotion analysis dataset in which Japanese social media posts
are annotated with the intensities of Plutchik's eight basic emotions,
and it is widely used for emotion classification of Japanese social media posts.
However, WRIME has an imbalanced label distribution across emotion categories,
and classification performance tends to be low for minority emotion labels
such as textit{Anger} and textit{Trust}.
In this study, we examine whether these minority labels
can be augmented by translating the English emotion-labeled dataset GoEmotions
into Japanese and adding the resulting pairs of Japanese sentences
and emotion labels to the training and development splits of WRIME.
For each English comment, we generated three Japanese translation candidates
and selected candidates based on BERTScore computed using back-translation,
cross-lingual cosine similarity with LaBSE,
and the original-label preservation probability estimated by a GoEmotions classifier.
In the classification experiments, we added a classification layer
to the T5 encoder and predicted the presence or absence of the eight emotions.
When trained only on WRIME, the F1 scores for textit{Anger} and textit{Trust}
were 0.3826 and 0.1897, respectively.
Adding GoEmotions-derived data only slightly improved
the Matthews correlation coefficient (MCC) for textit{Anger}
under some conditions and did not yield stable improvements for textit{Trust}.
These results suggest that
augmenting minority labels with external English emotion data requires
not only increasing the number of positive instances
but also ensuring consistency with the target dataset in terms of label definitions,
writing style, domain, and emotion intensity.
Keyword (in Japanese) (See Japanese page) 
(in English) emotion classification / label imbalance / data augmentation / machine translation / GoEmotions / WRIME / /  
Reference Info. IEICE Tech. Rep., vol. 126, no. 121, CQ2026-33, pp. 113-118, July 2026.
Paper # CQ2026-33 
Date of Issue 2026-07-14 (CQ) 
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 CQ2026-33

Conference Information
Committee CQ  
Conference Date 2026-07-21 - 2026-07-23 
Place (in Japanese) (See Japanese page) 
Place (in English) Eef Jouhou Plaza 
Topics (in Japanese) (See Japanese page) 
Topics (in English)  
Paper Information
Registration To CQ 
Conference Code 2026-07-CQ 
Language Japanese 
Title (in Japanese) (See Japanese page) 
Sub Title (in Japanese) (See Japanese page) 
Title (in English) Augmenting Minority Emotion Labels in Japanese Emotion Classification through Translation of English Emotion-Labeled Data: Effectiveness and Limitations in WRIME 
Sub Title (in English)  
Keyword(1) emotion classification  
Keyword(2) label imbalance  
Keyword(3) data augmentation  
Keyword(4) machine translation  
Keyword(5) GoEmotions  
Keyword(6) WRIME  
Keyword(7)  
Keyword(8)  
1st Author's Name Yuya Wake  
1st Author's Affiliation University of Tsukuba (Univ. of Tsukuba)
2nd Author's Name Sho Tsugawa  
2nd Author's Affiliation University of Tsukuba (Univ. of Tsukuba)
3rd Author's Name  
3rd Author's Affiliation ()
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 ()
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 2026-07-21 16:40:00 
Presentation Time 25 minutes 
Registration for CQ 
Paper # CQ2026-33 
Volume (vol) vol.126 
Number (no) no.121 
Page pp.113-118 
#Pages 6 
Date of Issue 2026-07-14 (CQ) 

[Return to Top Page]

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