检索结果-内蒙古大学图书馆

23rd International Conference on Neural Information Processing (ICONIP)

作者： Zhang, Li Lu, Yaping Zhang, Zhao Wang, Bangjun Li, Fanzhang Soochow Univ Sch Comp Sci & Technol Suzhou 215006 Jiangsu Peoples R China Soochow Univ Joint Int Res Lab Machine Learning & Neuromorph C Suzhou 215006 Jiangsu Peoples R China

ISBN: (纸本)9783319466750;9783319466743

To obtain a satisfying deep network, it is important to improve the performance on data representation of an auto-encoder. One of the strategies to enhance the performance is to incorporate sparsity into an auto-encoder. Fortunately, sparsity for the auto-encoder has been achieved by adding a Kullback-Leibler (KL) divergence term to the risk functional. In compressive sensing and machine learning, it is well known that the l(1) regularization is a widely used technique which can induce sparsity. Thus, this paper introduces a smoothed l(1) regularization instead of the mostly used KL divergence to enforce sparsity for auto-encoders. Experimental results show that the smoothed l(1) regularization works better than the KL divergence.

关键词： auto-encoder Sparsity KL divergence Smoothed l(1) regularization Data representation

来源：评论

学校读者我要写书评

暂无评论

An Intelligent Building Occupancy Detection System Based On Sparse auto-encoder 17

An Intelligent Building Occupancy Detection System Based On ...

引用

17th IEEE Winter Conference on Applications of Computer Vision (WACV)

作者： Liu, Zhi Zhang, Jie Geng, Li Chongqing Univ Technol Coll Comp Sci & Engn Chongqing Peoples R China Chongqing Univ Technol Coll Elect & Elect Engn Chongqing Peoples R China CUNY New York City Coll Technol New York NY 10021 USA

ISBN: (纸本)9781509049417

Energy saving in buildings has attracted many researchers attention. One of the research topics is occupancy detection for each room to automatically control air-conditioner, light, heating, etc. by monitoring the temperature, humidity, light, and co(2) using the corresponding sensors. For existing data analysis, traditional regression methods such as CART, RF, LDA are often used to predict the occupancy status after the data are collected. However, traditional methods have disadvantages of difficult adaptation to large data sets and bad generalization. In this paper, deep learning technology is used to perform occupancy detection. We construct sparse auto-encoder to learn features from data, then the learned features are input to Softmax, SVM and Liblinear classifiers to detect the room occupancy status. Experiments and results show the efficiency of the proposed algorithm whose detection accuracy outperforms that of benchmark project. Moreover, proposed algorithm has very low computation overhead which facilitates the real-time detection.

关键词： Occupancy Detection Deep Learning auto-encoder Sparse Coding Energy Saving

来源：评论

学校读者我要写书评

暂无评论

automatic Personality Perception Using auto-encoder And Hierarchical Fuzzy Classification 26

Automatic Personality Perception Using Auto-encoder And Hier...

引用

26th International Computer Conference of the Computer-Society-of-Iran

作者： Zaferani, Effat Jalaeian Teshnehlab, Mohammad Vali, Mansour KN Toosi Univ Technol Intelligent Syst Lab Elect & Comp Engn Fac Tehran Iran KN Toosi Univ Technol Speech Proc Lab Elect & Comp Engn Fac Tehran Iran

ISBN: (纸本)9781665412414

In this research, a study of automatic personality perception based on the Big-five Inventory (BFI) is done. To extract and select appropriate features for the classification, we employ an auto-encoder as a nonlinear feature learning technique. Since an auto-encoder does not extract proper classification lonely, a saddle point is found by a stop criterion based on maximum separate ability in binary classes. The results reveal that nonlinear features enhance the classification results in most personality traits. Furthermore, we use an adaptive neuro-fuzzy inference system classification to model the uncertainty rooted in mental states and affect the classification results through the extracted features. The classification outcomes on SSPNet Speaker Personality dataset demonstrate significant improvement in the results of four traits. These outgrowths verify the existence of uncertainty in the speech signal.

关键词： automatic Personality Perception Big-five Inventory Speech Signals auto-encoder Fuzzy Systems Feature Fusion

来源：评论

学校读者我要写书评

暂无评论

Binary Coding of Speech Spectrograms Using a Deep auto-encoder

Binary Coding of Speech Spectrograms Using a Deep Auto-encod...

引用

11th Annual Conference of the International-Speech-Communication-Association 2010

作者： Deng, L. Seltzer, M. Yu, D. Acero, A. Mohamed, A. Hinton, G. Microsoft Res 1 Microsoft Way Redmond WA 98052 USA Univ Toronto Toronto ON Canada

ISBN: (纸本)9781617821233

This paper reports our recent exploration of the layer-by-layer learning strategy for training a multi-layer generative model of patches of speech spectrograms. The top layer of the generative model learns binary codes that can be used for efficient compression of speech and could also be used for scalable speech recognition or rapid speech content retrieval. Each layer of the generative model is fully connected to the layer below and the weights on these connections are pre-trained efficiently by using the contrastive divergence approximation to the log likelihood gradient. After layer-by-layer pre-training we "unroll" the generative model to form a deep auto-encoder, whose parameters are then fine-tuned using back-propagation. To reconstruct the full-length speech spectrogram, individual spectrogram segments predicted by their respective binary codes are combined using an overlap-and-add method. Experimental results on speech spectrogram coding demonstrate that the binary codes produce a log-spectral distortion that is approximately 2 dB lower than a sub-band vector quantization technique over the entire frequency range of wide-band speech.

关键词： deep learning speech feature extraction neural networks auto-encoder binary codes Boltzmann machine

来源：评论

学校读者我要写书评

暂无评论

Application of Feature Selective auto-encoder in Aero Target Recognition

Application of Feature Selective Auto-encoder in Aero Target...

引用

2021 CIE International Conference on Radar, Radar 2021

作者： Deng, Sheng Guo, Xiaonan NO.38 Research Institute of Cetc Key Laboratory of Aperture Array and Space Application Hefei China

ISBN: (纸本)9781665498142

Rader aero target recognition has important application value on modern war field. In this paper, a feature selection auto-encoder based radar aero target classification method is proposed. The proposed method can automatically select manual features and further extract high-level features for classification of helicopter, propeller and jet. In experiments, real application data is used for the classification. Experimental results indicate the efficiency of the proposed method. © 2021 IEEE.

关键词： Aero target classification auto-encoder Deep learning Feature selection Radar

来源：评论

学校读者我要写书评

暂无评论

auto-encoder design based on the 1D-VD-CNN model for the detection of honeysuckle from unknown origin

引用

JOURNAL OF PHARMACEUTICAL AND BIOMEDICAL ANALYSIS 2023年 234卷 115572页

作者： Chen, Dongying Zhang, Hao Lin, Lingyan Zhang, Zilong Zeng, Jian Chen, Lu Chen, Xiaogang Fujian Jiangxia Univ Coll Elect Informat Sci Fuzhou 350108 Fujian Peoples R China Smart Home Informat Collect & Proc Internet Things Fuzhou 350108 Fujian Peoples R China Shandong Acad Agr Sci Inst Agr Qual Stand & Testing Technol Jinan 250100 Peoples R China Fuzhou Univ Coll Elect Engn & Automat Fuzhou 350108 Fujian Peoples R China

The disadvantages of the traditional one-dimensional convolution neural network (1D-CNN) model based on honeysuckle near-infrared spectral data (NIRS) include high parameter quantity, low efficiency, and inability to identify unknown categories effectively. In this paper, we propose a one-dimensional very deep convolution neural network (1D-VD-CNN) and design an auto-encoder mechanism for detecting honeysuckle from unexplored habitats. First, the 1D-VD-CNN model uses the efficient very deep (VD) structure to replace the hidden layer structure in the traditional 1D-CNN model. The model can be directly applied to analyze one-dimensional near-infrared spectral data (NIRS). Second, combining the reconstruction error of the auto-encoder, a honeysuckle identification method considering an unknown origin is designed, which can solve the problem of high confidence in convolution neural networks by using an auto-encoder and reconstruction errors of the samples to be tested. Whether the sample is an unknown variety can be determined by comparing the corrected confidence level with the preset threshold value. The results show that the accuracy of the 1D-VD-CNN training set and test set is 100%, and the loss value converges to 0.001. Compared with the traditional 1D-CNN model, the parameters and FLOPs are reduced by nearly 71% and 8%, respectively. At the same time, compared with the NIRS analysis and the PLS-DA method, the 1D-VD-CNN model has higher efficiency and better recognition performance for honeysuckle near-infrared spectral classification. Meanwhile, the accuracy rate of the auto-encoder for the category detection mechanism of honeysuckle from an unknown origin is 98%. The model can quickly and efficiently classify honeysuckle from different habitats and detect honeysuckle from unexplored habitats.

关键词： Honeysuckle NIRS auto-encoder 1D-VD-CNN Origin identification

来源：评论

学校读者我要写书评

暂无评论

Community detection in networks through a deep robust auto-encoder nonnegative matrix factorization

引用

ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE 2023年 118卷

作者： Al-sharoa, Esraa Rahahleh, Baraa Jordan Univ Sci & Technol Irbid 22110 Jordan

Community detection is a field of research with increasing influence due to its importance in dimensionality reduction and revealing the organizational patterns in networks. A community or cluster in a network reflects the interactions among the objects that are strongly connected and distinguish them from the ones in other clusters. Over the past decades, various algorithms have been developed to detect the community structure in networks. However, the performance of these algorithm is affected by the existence of noise or nonlinear information in the network. Consequently, deep learning algorithms arise to tackle this issue. In this paper, we propose a novel deep robust auto-encoder nonnegative matrix factorization (DRANMF) approach to detect the community structure in networks. In particular, DRANMF consists of a deep structured decoder and encoder components to transform the high-order proximity matrix of the network to and from the cluster membership space, respectively. Moreover, we enhance the robustness of the proposed approach against noise and outliers by adopting the I-2,I-1-norm to formulate the objective function. The experiments on multiple synthetic and real-world networks demonstrate that the proposed DRANMF has a better performance and robustness than other existing methods.

关键词： Networks auto-encoder Deep learning Community detection Nonnegative matrix factorization

来源：评论

学校读者我要写书评

暂无评论

An auto-encoder based LSTM model for prediction of ambient noise levels

引用

APPLIED ACOUSTICS 2022年 195卷

作者： Tiwari, S. K. Kumaraswamidhas, L. A. Gautam, C. Garg, N. Indian Inst Technol ISM Dhanbad 826 001 India CSIR Natl Phys Lab New Delhi 110 012 India

Traffic noise is one of the most prevalent cause of environmental pollution in Indian cities. A reliable method is required for assessment, and prediction of ambient noise levels. This paper presents a novel deep learning model based on auto-encoder infused with Long short-term memory (LSTM), to predict ambient noise levels. The model automatically selects the best prediction technique by considering dif-ferent combination of hyper-parameters using grid search methodology. It has the ability to inherit non-stationary characteristics of time-series data while considering non-linear pattern. The proposed model is compared with some well-known techniques like Artificial neural technique (ANN), Support vector machine (SVM), Recurrent neural network (RNN), and Long short term memory (LSTM) model. The study concludes that the proposed model outperforms other techniques and can be a reliable approach for time-series prediction of ambient noise levels with an error of +/- 0.563 dB(A). The prediction capability of the models is ascertained by statistical tests parameters namely RMSE, MAE, R-2, and ACC% which is fur-ther validated by Friedman test.(C) 2022 Elsevier Ltd. All rights reserved.

关键词： Deep learning auto-encoder LSTM Data preprocessing Noise level prediction Friedman test

来源：评论

学校读者我要写书评

暂无评论

Frequency Selective auto-encoder for Smart Meter Data Compression

引用

SENSORS 2021年第4期21卷 1521-1521页

作者： Lee, Jihoon Yoon, Seungwook Hwang, Euiseok Gwangju Inst Sci & Technol GIST Sch Elect Engn & Comp Sci 123 Cheomdangwagi Ro Gwangju 61005 South Korea Gwangju Inst Sci & Technol GIST Sch Mechatron 123 Cheomdangwagi Ro Gwangju 61005 South Korea

With the development of the internet of things (IoT), the power grid has become intelligent using massive IoT sensors, such as smart meters. Generally, installed smart meters can collect large amounts of data to improve grid visibility and situational awareness. However, the limited storage and communication capacities can restrain their infrastructure in the IoT environment. To alleviate these problems, efficient and various compression techniques are required. Deep learning-based compression techniques such as auto-encoders (AEs) have recently been deployed for this purpose. However, the compression performance of the existing models can be limited when the spectral properties of high-frequency sampled power data are widely varying over time. This paper proposes an AE compression model, based on a frequency selection method, which improves the reconstruction quality while maintaining the compression ratio (CR). For efficient data compression, the proposed method selectively applies customized compression models, depending on the spectral properties of the corresponding time windows. The framework of the proposed method involves two primary steps: (i) division of the power data into a series of time windows with specified spectral properties (high-frequency, medium-frequency, and low-frequency dominance) and (ii) separate training and selective application of the AE models, which prepares them for the power data compression that best suits the characteristics of each frequency. In simulations on the Dutch residential energy dataset, the frequency-selective AE model shows significantly higher reconstruction performance than the existing model with the same CR. In addition, the proposed model reduces the computational complexity involved in the analysis of the learning process.

关键词： data compression smart meter auto-encoder digital signal processing

来源：评论

学校读者我要写书评

暂无评论

Multi-kernel fuzzy clustering based on auto-encoder for fMRI functional network

引用

EXPERT SYSTEMS WITH APPLICATIONS 2020年 159卷 113513-113513页

作者： Lu, Hu Liu, Saixiong Wei, Hui Tu, Juanjuan Jiangsu Univ Sch Comp Sci & Commun Engn 301 Xuefu Rd Zhenjiang 212013 Jiangsu Peoples R China Fudan Univ Sch Comp Sci Lab Cognit & Model Algorithm Shanghai 200433 Peoples R China Jiangsu Univ Sci & Technol Sch Comp Sci Zhenjiang 212003 Jiangsu Peoples R China

The existing clustering algorithms based on auto-encoder only use one layer of information. In this paper, we propose a new subspace clustering algorithm that uses multiple hidden layers of information from the stacked auto-encoder to construct different kernels. The proposed fuzzy multi-kernel clustering method based on auto-encoder is realized by updating the membership matrix and coefficients of these kernels until the value of objective function is iterated to the minimum error. Also, the proposed method combining auto-encoder achieves high dimension reduction of input data. In order to test the effectiveness of the algorithm, we first performed experiments on the published brain network dataset. Compared with MKFC, RMKKM and other algorithms, the proposed method can significantly improve accuracy. Specifically, the brain network has large dimensions when the functional magnetic resonance imaging (fMRI) data is preprocessed. Therefore, the method in this paper is applied to the Autism Spectrum Disorder (ASD) of the Autism Brain Imaging Data Exchange (ABIDE) database. The experiment results on the network dataset we constructed with high dimension are better than the present several clustering algorithms. The results show that the subspace information after dimensionality reduction is more conducive to clustering. (C) 2020 Elsevier Ltd. All rights reserved.

关键词： Clustering Multi-kernel auto-encoder Functional network

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：