Deep Learning of Part-Based Representation of Data Using Sparse Autoencoders With Nonnegativity Constraints

Ehsan Hosseini-Asl; Jacek M. Zurada; Olfa Nasraoui

首页> 外文期刊>Neural Networks and Learning Systems, IEEE Transactions on >Deep Learning of Part-Based Representation of Data Using Sparse Autoencoders With Nonnegativity Constraints

【24h】

Deep Learning of Part-Based Representation of Data Using Sparse Autoencoders With Nonnegativity Constraints

机译：使用具有非负约束的稀疏自动编码器对基于零件的数据表示进行深度学习

获取原文

获取原文并翻译 | 示例

掌桥外文数据库（机构版） >>

开具论文收录证明 >>

文献代查 >>

页面导航

摘要
著录项
相似文献
相关主题

摘要

We demonstrate a new deep learning autoencoder network, trained by a nonnegativity constraint algorithm (nonnegativity-constrained autoencoder), that learns features that show part-based representation of data. The learning algorithm is based on constraining negative weights. The performance of the algorithm is assessed based on decomposing data into parts and its prediction performance is tested on three standard image data sets and one text data set. The results indicate that the nonnegativity constraint forces the autoencoder to learn features that amount to a part-based representation of data, while improving sparsity and reconstruction quality in comparison with the traditional sparse autoencoder and nonnegative matrix factorization. It is also shown that this newly acquired representation improves the prediction performance of a deep neural network.

机译：我们演示了一种新的深度学习自动编码器网络，该网络由非负约束算法（非负约束自动编码器）训练，可以学习显示基于部分数据表示的特征。学习算法基于约束负权重。基于将数据分解为多个部分来评估算法的性能，并在三个标准图像数据集和一个文本数据集上测试其预测性能。结果表明，与传统的稀疏自动编码器和非负矩阵分解相比，非负约束迫使自动编码器学习相当于基于部分数据表示的特征，同时提高了稀疏性和重构质量。还表明，这种新获得的表示提高了深度神经网络的预测性能。

著录项

来源
《Neural Networks and Learning Systems, IEEE Transactions on》 |2016年第12期|2486-2498|共13页
作者
Ehsan Hosseini-Asl; Jacek M. Zurada; Olfa Nasraoui;
展开▼
作者单位

Department of Electrical and Computer Engineering, University of Louisville, Louisville, KY, USA;

Department of Electrical and Computer Engineering, University of Louisville, Louisville, KY, USA;

Department of Computer Science, University of Louisville, Louisville, KY, USA;

展开▼
收录信息
原文格式 PDF
正文语种 eng
中图分类
关键词
Training; Feature extraction; Artificial neural networks; Machine learning; Image reconstruction; Encoding; Cost function;

机译：训练;特征提取;人工神经网络;机器学习;图像重建;编码;成本函数;

相似文献

外文文献
中文文献
专利

1. Automatic Modulation Classification Using Deep Learning Based on Sparse Autoencoders With Nonnegativity Constraints [J] . Afan Ali, Fan Yangyu IEEE signal processing letters . 2017,第11期

机译：基于具有非负约束的稀疏自编码器的深度学习自动调制分类
2. Learning a referenceless stereopair quality engine with deep nonnegativity constrained sparse autoencoder [J] . Jiang Qiuping, Shao Feng, Lin Weisi, Pattern Recognition: The Journal of the Pattern Recognition Society . 2018,第期

机译：学习具有深度非室内基间的引用立体声质量引擎约束稀疏自动化器
3. Dictionary Learning with Incoherence and Sparsity Constraints for Sparse Representation of Nonnegative Signals [J] . Zunyi TANG, Shuxue DING IEICE transactions on information and systems . 2013,第5期

机译：具有非相干性和稀疏性约束的字典学习用于非负信号的稀疏表示
4. Visualizing and Understanding Nonnegativity Constrained Sparse Autoencoder in Deep Learning [C] . Babajide O. Ayinde, Ehsan Hosseini-Asl, Jacek M. Zurada International conference on artificial intelligence and soft computing . 2016

机译：可视化和理解深度学习中的非负约束稀疏自动编码器
5. Sparse Representation and Deep Learning for Image and Video Reconstruction. [D] . Iliadis, Michael. 2016

机译：图像和视频重建的稀疏表示和深度学习。
6. Subspace structural constraint-based discriminative feature learning via nonnegative low rank representation [O] . Ao Li, Xin Liu, Yanbing Wang, 2015

机译：通过非负低秩表示的基于子空间结构约束的判别特征学习
7. Deep Learning of Part-based Representation of Data Using Sparse Autoencoders with Nonnegativity Constraints [O] . Hosseini-Asl, Ehsan, Zurada, Jacek M., Nasraoui, Olfa 2016

机译：利用稀疏差分深度学习基于部分的数据表示具有非负性约束的自动编码器

Deep Learning of Part-Based Representation of Data Using Sparse Autoencoders With Nonnegativity Constraints

摘要

著录项

相似文献

相关主题

期刊订阅