Discriminative Auto-Encoder With Local and Global Graph Embedding
In order to exploit the potential intrinsic low-dimensional structure of the high-dimensional data from the manifold learning perspective, we propose a global graph embedding with globality-preserving property, which requires that samples should be mapped close to their low-dimensional class represe...
Main Authors: | , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
IEEE
2020-01-01
|
Series: | IEEE Access |
Subjects: | |
Online Access: | https://ieeexplore.ieee.org/document/8985268/ |
id |
doaj-c7c266826a0e438f84ebc8707806126a |
---|---|
record_format |
Article |
spelling |
doaj-c7c266826a0e438f84ebc8707806126a2021-03-30T02:08:14ZengIEEEIEEE Access2169-35362020-01-018286142862310.1109/ACCESS.2020.29721328985268Discriminative Auto-Encoder With Local and Global Graph EmbeddingRui Li0https://orcid.org/0000-0003-1770-906XXiaodan Wang1https://orcid.org/0000-0003-2785-9539Jie Lai2https://orcid.org/0000-0002-1028-8162Yafei Song3https://orcid.org/0000-0003-0962-0671Lei Lei4https://orcid.org/0000-0001-8415-8611College of Air and Missile Defense, Air Force Engineering University, Xi’an, ChinaCollege of Air and Missile Defense, Air Force Engineering University, Xi’an, ChinaCollege of Air and Missile Defense, Air Force Engineering University, Xi’an, ChinaCollege of Air and Missile Defense, Air Force Engineering University, Xi’an, ChinaCollege of Air and Missile Defense, Air Force Engineering University, Xi’an, ChinaIn order to exploit the potential intrinsic low-dimensional structure of the high-dimensional data from the manifold learning perspective, we propose a global graph embedding with globality-preserving property, which requires that samples should be mapped close to their low-dimensional class representation data distribution centers in the embedding space. Then we propose a novel local and global graph embedding auto-encoder(LGAE) to capture the geometric structure of data, its cost function have three terms, a reconstruction loss to reproduce the input data based on the learned representation, a local graph embedding regularization to enforce mapping the neighboring samples close together in the embedding space, a global embedding regularization to enforce mapping samples close to their low-dimensional class representation distribution centers. Thus in the learning process, our LGAE can map samples from same class close together in the embedding space, as well as reduce the scatter within-class and increase the margin between-class, it will also detect the local and global intrinsic geometric structure of data and discover the latent discriminant information in the embedding space. We build stacked LGAE for classification tasks and conduct comprehensive experiments on several benchmark datasets, the results confirm that our proposed framework can learn discriminative representation, speed up the network convergence process, and significantly improve the classification performance.https://ieeexplore.ieee.org/document/8985268/Manifold learninggraph embeddingdiscriminative auto-encoderdeep learning |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Rui Li Xiaodan Wang Jie Lai Yafei Song Lei Lei |
spellingShingle |
Rui Li Xiaodan Wang Jie Lai Yafei Song Lei Lei Discriminative Auto-Encoder With Local and Global Graph Embedding IEEE Access Manifold learning graph embedding discriminative auto-encoder deep learning |
author_facet |
Rui Li Xiaodan Wang Jie Lai Yafei Song Lei Lei |
author_sort |
Rui Li |
title |
Discriminative Auto-Encoder With Local and Global Graph Embedding |
title_short |
Discriminative Auto-Encoder With Local and Global Graph Embedding |
title_full |
Discriminative Auto-Encoder With Local and Global Graph Embedding |
title_fullStr |
Discriminative Auto-Encoder With Local and Global Graph Embedding |
title_full_unstemmed |
Discriminative Auto-Encoder With Local and Global Graph Embedding |
title_sort |
discriminative auto-encoder with local and global graph embedding |
publisher |
IEEE |
series |
IEEE Access |
issn |
2169-3536 |
publishDate |
2020-01-01 |
description |
In order to exploit the potential intrinsic low-dimensional structure of the high-dimensional data from the manifold learning perspective, we propose a global graph embedding with globality-preserving property, which requires that samples should be mapped close to their low-dimensional class representation data distribution centers in the embedding space. Then we propose a novel local and global graph embedding auto-encoder(LGAE) to capture the geometric structure of data, its cost function have three terms, a reconstruction loss to reproduce the input data based on the learned representation, a local graph embedding regularization to enforce mapping the neighboring samples close together in the embedding space, a global embedding regularization to enforce mapping samples close to their low-dimensional class representation distribution centers. Thus in the learning process, our LGAE can map samples from same class close together in the embedding space, as well as reduce the scatter within-class and increase the margin between-class, it will also detect the local and global intrinsic geometric structure of data and discover the latent discriminant information in the embedding space. We build stacked LGAE for classification tasks and conduct comprehensive experiments on several benchmark datasets, the results confirm that our proposed framework can learn discriminative representation, speed up the network convergence process, and significantly improve the classification performance. |
topic |
Manifold learning graph embedding discriminative auto-encoder deep learning |
url |
https://ieeexplore.ieee.org/document/8985268/ |
work_keys_str_mv |
AT ruili discriminativeautoencoderwithlocalandglobalgraphembedding AT xiaodanwang discriminativeautoencoderwithlocalandglobalgraphembedding AT jielai discriminativeautoencoderwithlocalandglobalgraphembedding AT yafeisong discriminativeautoencoderwithlocalandglobalgraphembedding AT leilei discriminativeautoencoderwithlocalandglobalgraphembedding |
_version_ |
1724185781930557440 |