.png){kind=icon}
返回
Multilayer discriminative extreme learning machine for classification
.png){kind=icon}
PRE
AI
DOI：10.1007/s13042-022-01749-7.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
The representation learning is the key to deep learning. As a special deep learning algorithm, the generalization performance of the multilayer extreme learning machine (ML-ELM) is influenced by the feature extraction capability of the extreme learning machine autoencoder (ELM-AE). But the ELM-AE does not consider class labels, and it is difficult for the ELM-AE to build the discriminative feature space in the classification tasks. Thus, to improve the class separability of abstract feature, the discriminant regularization is introduced into the ELM-AE and a new regularized ELM-AE called discriminative extreme learning machine autoencoder (DELM-AE) is proposed. And by stacking the ELM-AE for feature extraction and adopting the ELM for classifying, the multilayer discriminative extreme learning machine (ML-DELM) is proposed for classification tasks. The DELM-AE adds the discriminant regularization term to the loss function to reduce the distance between samples and their class centers in feature space. Compared with the ELM-AE and ML-ELM, the DELM-AE and ML-DELM indirectly utilizes the category information contained in the discriminant regularization term to guide the feature extraction. Empirical evaluations and experiments on various benchmark datasets show that feature representation learned by the DELM-AE is discriminative and sparse, and the generalization performance and structural sparsity of ML-DELM are comparable with the other state-of-the-art ML-ELM algorithms.
Keyword:
Extreme learning machine
Deep learning
Representation learning
Autoencoder
Discriminant regularization
期刊
IF:
2.7
论文数:
3.2K
被引数:
5.6K
机构
引用论文
Productivity enhancement of solar still by PCM and Nanoparticles miscellaneous basin absorbing materials.png)
Desalination
OAAI
DesalinationIF0
OAAI分享收藏Sheltering effect of vertical twin tunnel excavation for different horizontal distances on the ground and pipelinesSādhanā
SādhanāIF0
PREAI
分享收藏Hierarchical extreme learning machine with L21-norm loss and regularization具有L21-norm损失和正则化的分层极限学习机
PREAI
分享收藏Generalized extreme learning machine autoencoder and a new deep neural networkNEUROCOMPUTING
NEUROCOMPUTINGIF6.5
PREAI
分享收藏L21-Norm Based Loss Function and Regularization Extreme Learning Machine基于L21-Norm损失函数和正则化的极限学习机IEEE ACCESSOAAI
IEEE ACCESSIF3.6
OAAI分享收藏
查看更多内容