检索结果-内蒙古大学图书馆

56th Asilomar Conference on Signals, Systems, and Computers

作者： Hasssan, Ahmed Meng, Jian Cao, Yu Seo, Jae-sun Arizona State Univ Sch Elect Comp & Energy Engn Tempe AZ 85281 USA

ISBN: (纸本)9781665459068

Imaging innovations such as dynamic vision sensor (DVS) can significantly reduce the image data volume by tracking only the changes in events. However, when DVS camera itself moves around (e.g. self-driving cars), the DVS output stream is not sparse enough to achieve the desired hardware efficiency. In this work, we investigate designing a compact sparse autoencoder model to largely compress event-based DVS output. The proposed encoder-decoder-based autoencoder design is a shallow convolutional neural network (CNN) architecture with two convolution and inverse-convolution layers with only similar to 10k parameters. We implement quantization-aware training on our proposed model to achieve 2-bit and 4-bit precision. Moreover, we implement unstructured pruning on the encoder module to achieve >90% active pixel compression at the latent space. The proposed autoencoder design has been validated against multiple benchmark DVS-based datasets including DVS-MNIST, N-Cars, DVS-IBM Gesture, and Prophesee Automotive Gen1 dataset. We achieve low accuracy drop of 2%, 3%, and 3.8% compared to the uncompressed baseline, with 7.08%, 1.36%, and 5.53% active pixels in the images from the decoder (compression ratio of 12.1x, 29.1x, and 18.1x) for DVS-MNIST, N-Cars, and DVS-IBM Gesture datasets, respectively. For the Prophesee Automotive Gen1 dataset, we achieve a minimal mAP drop of 0.07 from the baseline with 9% active pixels in the images from the decoder (compression ratio of 11.9x).

关键词： Neural network training Dynamic vision sensor sparse autoencoder Neural network pruning Quantization Object recognition detection

来源：评论

学校读者我要写书评

暂无评论

Transformer Fault Diagnosis based on Deep Brief sparse autoencoder 38

Transformer Fault Diagnosis based on Deep Brief Sparse Autoe...

引用

38th Chinese Control Conference (CCC)

作者： Xu, Zhong Mo, Wenxiong Wang, Yong Luo, Simin Liu, Tian Guangzhou Power Supply Bur Co Ltd Elect Power Test & Res Inst Guangzhou 510000 Peoples R China

ISBN: (纸本)9789881563972

Dissolved gas analysis (DGA) is an effective way to diagnose the internal faults of transformer. This paper proposes a deep belief sparse autoencoder (DBSAE), which can be performed on DGA data to detect the transformer fault. The proposed method is constructed by two layers of sparse autoencoder and one layer of back propagation network. DBSAE adopts the unsupervised training to collect the features of DGA raw data, and the back propagation network with the supervised training could realize the transformer faults diagnosis. The experiments are performed on a real DGA dataset. The experimental results verify the effectiveness of the proposed method, which can extract features from the original data, and achieve a superior correct differentiation rate which reached 95.4% in average accuracy on transformer fault diagnosis.

关键词： Transformer fault diagnosis dissolved gas analysis sparse autoencoder deep belief network

来源：评论

学校读者我要写书评

暂无评论

A sparse autoencoder compressed sensing method for acquiring the pressure array information of clothing

引用

Neurocomputing 2018年 275卷 1500-1510页

作者： Han, Tao Hao, Kuangrong Ding, Yongsheng Tang, Xuesong Engineering Research Center of Digitized Textile & Apparel Technology Ministry of Education Donghua University Shanghai 201620 China College of Information Sciences and Technology Donghua University 2999 Renmin North Road Songjiang District Shanghai 201620 China

In this paper, we integrate some ideas of sparse autoencoder of deep learning into compressed sensing (CS) theory, and set up a sparse autoencoder compressed sensing (SAECS) model, which can improve the compressed sampling process of CS with compression of sparse autoencoder in deep learning. The original CS theory has no function of autonomic regulation, so we introduce the idea of sparse autoencoder of deep learning to improve CS theory. Then we calculate the error between the recovery output data and the input data. By judging the obtained error and the acceptable error, the SAECS model can choose autonomously the most appropriate sparsity and the most appropriate length of measurement vector. This SAECS model can then reconstruct the original signal that can satisfy the acceptable error requirement with the minimum length of measurement vector in CS theory. We investigate the effectiveness of the proposed method by using sampled pressure data from human body model. Experimental results demonstrate that, in comparison with the state-of-the-art methods for running time, our SACES approach can effectively decrease the running time to find the shortest measurement vector in the case of accepted error. © 2017 Elsevier B.V.

关键词： Compressed sensing Deep learning Human body Neural network Pressure array sparse autoencoder

来源：评论

学校读者我要写书评

暂无评论

Image edge detection based on sparse autoencoder network 10

Image edge detection based on Sparse Autoencoder network

引用

10th International Conference on Information Optics and Photonics

作者： Liu, Yingwei Gao, Xiaorong Li, Jinlong Southwest Jiaotong Univ Sch Phys Sci & Technol 111 Sect 1 North Ring Rd Chengdu Sichuan Peoples R China

ISBN: (纸本)9781510625808

Edge detection plays an important role in image pattern recognition. Because of the shortcomings of poor anti-noise and spurious edges by using traditional edge detection methods. A method of image edge detection based on sparse autoencoder neural work is proposed in this paper. This method uses Berkeley Segmentation data set to extract the high-dimensional edge features of sample data by training the sparse autoencoder. Through the ZCA (Zero-phase Component Analysis) whitening treatment, the correlation between images is effectively reduced. The standard edge images are input into a Softmax classifier to train a classifier that can classify the edge features of each pixel. Last, the extracted features of each pixel sample are input into the trained Softmax classifier to classify the edge pixels to achieve edge detection. Experiments show that the algorithm has good noise immunity and certain application value.

关键词： Edge detection sparse autoencoder ZCA whitening Softmax Pattern recognition

来源：评论

学校读者我要写书评

暂无评论

PREPROCESSING-FREE SURFACE MATERIAL CLASSIFICATION USING CONVOLUTIONAL NEURAL NETWORKS PRETRAINED BY sparse autoencoder 25

PREPROCESSING-FREE SURFACE MATERIAL CLASSIFICATION USING CON...

引用

IEEE International Workshop on Machine Learning for Signal Processing

作者： Ji, M. Q. Fang, Lu Zheng, Haitian Strese, Matti Steinbach, Eckehard Hong Kong Univ Sci & Technol Hong Kong Hong Kong Peoples R China Univ Sci & Technol China Hefei Anhui Peoples R China Tech Univ Munich D-80290 Munich Germany

ISBN: (纸本)9781467374545

Acceleration signals captured during the interaction of a rigid tool with an object surface carry relevant information for surface material classification. Existing methods mostly rely on carefully designed perception-related features or features adapted from audio processing motivated by the observed similarity between acceleration signals and audio signals. In contrast, our proposed method automatically learns features from RAW acceleration data without preprocessing. The approach is based on Convolutional Neural Networks (CNN) trained and tested on RAW data. For better performance and faster convergence of the CNN, we use the weights of a trained sparse autoencoder (AE) to initialize the weights of the first convolution layers of the CNN. This strategy is named CNN pretrained by sparse AE (ACNN). Our classification results on a publically available Haptic Texture Database demonstrate that the proposed algorithm performs favorably against existing methods.

关键词： haptic texture classification convolutional neural networks CNN pretraining sparse autoencoder

来源：评论

学校读者我要写书评

暂无评论

Anomaly Detection in Network Traffic Using Dynamic Graph Mining with a sparse autoencoder 18

Anomaly Detection in Network Traffic Using Dynamic Graph Min...

引用

18th IEEE International Conference on Trust, Security and Privacy in Computing and Communications (IEEE TrustCom) / 13th IEEE International Conference on Big Data Science and Engineering (IEEE BigDataSE)

作者： Jia, Guanbo Miller, Paul Hong, Xin Kalutarage, Harsha Ban, Tao Queens Univ Belfast Ctr Secur Informat Technol Belfast Antrim North Ireland Robert Gordon Univ Sch Comp Sci & Digital Media Aberdeen Scotland Natl Inst Informat & Commun Technol Informat Secur Res Ctr Koganei Tokyo Japan

ISBN: (纸本)9781728127767

Network based attacks on ecommerce websites can have serious economic consequences. Hence, anomaly detection in dynamic network traffic has become an increasingly important research topic in recent years. This paper proposes a novel dynamic Graph and sparse autoencoder based Anomaly Detection algorithm named GAAD. In GAAD, the network traffic over contiguous time intervals is first modelled as a series of dynamic bipartite graph increments. One mode projection is performed on each bipartite graph increment and the adjacency matrix derived. Columns of the resultant adjacency matrix are then used to train a sparse autoencoder to reconstruct it. The sum of squared errors between the reconstructed approximation and original adjacency matrix is then calculated. An online learning algorithm is then used to estimate a Gaussian distribution that models the error distribution. Outlier error values are deemed to represent anomalous traffic flows corresponding to possible attacks. In the experiment, a network emulator was used to generate representative ecommerce traffic flows over a time period of 225 minutes with five attacks injected, including SYN scans, host emulation and DDoS attacks. ROC curves were generated to investigate the influence of the autoencoder hyper-parameters. It was found that increasing the number of hidden nodes and their activation level, and increasing sparseness resulted in improved performance. Analysis showed that the sparse autoencoder was unable to encode the highly structured adjacency matrix structures associated with attacks, hence they were detected as anomalies. In contrast, SVD and variants, such as the compact matrix decomposition, were found to accurately encode the attack matrices, hence they went undetected.

关键词： anomaly detection network traffic network security bipartite graph sparse autoencoder dynamic graph

来源：评论

学校读者我要写书评

暂无评论

Blade Imbalanced Fault Diagnosis for Marine Current Turbine based on sparse autoencoder and Softmax Regression 33

Blade Imbalanced Fault Diagnosis for Marine Current Turbine ...

引用

33rd Youth Academic Annual Conference of Chinese Association of Automation (YAC)

作者： Wen, Pingping Wang, Tianzhen Xin, Bin Tang, Tianhao Wang, Yide Shanghai Maritime Univ Dept Elect Automat Shanghai Peoples R China Nantes Univ Inst Elect & Elect Engn Nantes France

ISBN: (纸本)9781538672556

Because of the abundance of seston under the sea, the attachment on the blade of the marine current turbine (MCT) would cause imbalanced fault. In order to detect the imbalanced fault as soon as possible, an imbalanced fault characteristics analysis method is applied based on image processing. A diagnosis method combining the modified sparse autoencoder (SA) and softmax regression (SR) is applied to process images and detect the imbalanced fault on the blade of MCT. The modified SA is used to extract the features and SR is used to classify them. The data of images are used to monitor whether the blade is attached by benthos and its corresponding degree of imbalance. Experiments show that the applied diagnosis method can achieve higher accuracy in the application of diagnosis of blade imbalanced fault compared with the traditional PCA feature extraction algorithm.

关键词： marine current turbine blade imbalanced fault sparse autoencoder softmax regression

来源：评论

学校读者我要写书评

暂无评论

Pseudoinverse Learning Algorithom for Fast sparse autoencoder Training

Pseudoinverse Learning Algorithom for Fast Sparse Autoencode...

引用

IEEE Congress on Evolutionary Computation (IEEE CEC) as part of the IEEE World Congress on Computational Intelligence (IEEE WCCI)

作者： Xu, Bingxin Guo, Ping Beijing Union Univ Beijing Key Lab Informat Serv Engn Beijing 100101 Peoples R China Beijing Normal Univ Sch Syst Sci Image Proc & Pattern Recognit Lab Beijing 100875 Peoples R China

ISBN: (纸本)9781509060177

sparse autoencoder is one approach to automatically learn features from unlabeled data and received significant attention during the development of deep neural networks. However, the learning algorithm of sparse autoencoder suffers from slow learning speed because of gradient descent based algorithms have many drawbacks. In this paper, a fast learning algorithm for sparse autoenceder is proposed which based on pseudoinverse learning algorithm (PIL). The proposed method calculates encoder weight matrix by truncating the pseudoinverse matrix of input data. The pseudoinverse truncation matrix is used as the weights of encoder, and then the input data is mapped to the hidden layer space through the biased ReLU activation function. The decoder weights are also can computed by the PIL. Unlike the gradient descent based algorithm, the proposed method does not require a time-consuming iterative optimization process and select many user-dependent parameters such as learning rate or momentum constant too. The experimental results indicate the superiority of proposed method which is very efficient and also can learned the sparsity of samples.

关键词： sparse autoencoder pseudoinverse learning algorithm fast learning data representation

来源：评论

学校读者我要写书评

暂无评论

Extracting Activity Patterns from Altering Biological Networks: a sparse autoencoder Approach 32

Extracting Activity Patterns from Altering Biological Networ...

引用

32nd IEEE International Symposium on Computer-Based Medical Systems (IEEE CBMS)

作者： Szlobodnyik, Gergely Pazmany Peter Catholic Univ Dept Informat Technol & Bion Budapest Hungary

ISBN: (纸本)9781728122861

In this paper we address the problem of extracting activity patterns from biological networks that cannot be characterized in the form of a static graph, such as gene regulatory networks wherein the co-expression patterns of different genes may vary over time. A simple autoencoder architecture is applied in order to detect activity patterns appearing in several observational samples of the data. We apply Kullback-Leibler divergence based sparsity constraint on the hidden layer neurons, such a way the activity of a hidden layer unit can be associated to a particular activity pattern which can be extracted using the trained decoder function.

关键词： activity patterns network inference altering graph neural networks sparse autoencoder

来源：评论

学校读者我要写书评

暂无评论

Face recognition based on deep aggregated sparse autoencoder network 37

Face recognition based on deep aggregated sparse autoencoder...

引用

37th Chinese Control Conference (CCC)

作者： Zou, Guofeng Lin, Dingyi Fu, Gui-xia Shen, Jin Gao, Mingliang Shandong Univ Technol Coll Elect & Elect Engn Zibo 255049 Peoples R China

ISBN: (纸本)9789881563958

sparse autoencoder network is sensitive to face noise, and the learning process is easy to ignore the face structure information. Address this problem, we propose a face recognition approach fused sub-region LBP feature and deep aggregated sparse autoencoder network. Firstly, the face image is divided into different sub-regions, and the local binary pattern is used to pre-process the face image in order to obtain the LBP features of different sub-region faces. Then, the deep features for different sub-region LBP features are extracted with different sparse autoencoders. Finally, the output features of different sparse autoencoders are aggregated by full connection, and the face feature vectors are obtained for classification. We get the optimal aggregated network structure and network parameters through lots of experiments, while the face recognition accuracy is also improved.

关键词： Unsupervised deep learning sparse autoencoder deep aggregated network LBP feature sub-region division face recognition

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：