HUSCAP logo Hokkaido Univ. logo

Hokkaido University Collection of Scholarly and Academic Papers >
Office of Institutional Research >
Peer-reviewed Journal Articles, etc >

Multi-Task Convolutional Neural Network Leading to High Performance and Interpretability via Attribute Estimation

Files in This Item:
2020-12-01_Multi-ta.pdf930.68 kBPDFView/Open
Please use this identifier to cite or link to this item:http://hdl.handle.net/2115/80378

Title: Multi-Task Convolutional Neural Network Leading to High Performance and Interpretability via Attribute Estimation
Authors: Maeda, Keisuke Browse this author
Horii, Kazaha Browse this author
Ogawa, Takahiro Browse this author →KAKEN DB
Haseyama, Miki Browse this author →KAKEN DB
Keywords: multi-task convolutional neural network
image classification
attribute estimation
interpretability
Issue Date: Dec-2020
Publisher: 電子情報通信学会(The Institute of Electronics, Information and Communication Engineers / IEICE)
Journal Title: IEICE transactions on fundamentals of electronics communications and computer sciences
Volume: E103A
Issue: 12
Start Page: 1609
End Page: 1612
Publisher DOI: 10.1587/transfun.2020SML0006
Abstract: A multi-task convolutional neural network leading to high performance and interpretability via attribute estimation is presented in this letter. Our method can provide interpretation of the classification results of CNNs by outputting attributes that explain elements of objects as a judgement reason of CNNs in the middle layer. Furthermore, the proposed network uses the estimated attributes for the following prediction of classes. Consequently, construction of a novel multi-task CNN with improvements in both of the interpretability and classification performance is realized.
Rights: copyright©2020 IEICE
http://search.ieice.org/
Type: article
URI: http://hdl.handle.net/2115/80378
Appears in Collections:総合IR室 (Office of Institutional Research) > 雑誌発表論文等 (Peer-reviewed Journal Articles, etc)

Submitter: 前田 圭介

Export metadata:

OAI-PMH ( junii2 , jpcoar_1.0 )

MathJax is now OFF:


 

 - Hokkaido University