检索结果-内蒙古大学图书馆

CDEC:a constrained deep embedded clustering

International Journal of Intelligent Computing and Cybernetics 2021年第4期14卷 686-701页

作者： Elham Amirizadeh Reza Boostani Department of Computer Sciences and Engineering Shiraz UniversityShirazIran

Purpose-The aim of this study is to propose a deep neural network(DNN)method that uses side information to improve clustering results for big datasets;also,the authors show that applying this information improves the performance of clustering and also increase the speed of the network training ***/methodology/approach-In data mining,semisupervised learning is an interesting approach because good performance can be achieved with a small subset of labeled data;one reason is that the data labeling is expensive,and semisupervised learning does not need all *** type of semisupervised learning is constrained clustering;this type of learning does not use class labels for ***,it uses information of some pairs of instances(side information),and these instances maybe are in the same cluster(must-link[ML])or in different clusters(cannot-link[CL]).Constrained clustering was studied extensively;however,little works have focused on constrained clustering for big *** this paper,the authors have presented a constrained clustering for big datasets,and the method uses a *** authors inject the constraints(ML and CL)to this DNN to promote the clustering performance and call it constrained deep embedded clustering(CDEC).In this manner,an autoencoder was implemented to elicit informative low dimensional features in the latent space and then retrain the encoder network using a proposed Kullback-Leibler divergence objective function,which captures the constraints in order to cluster the projected *** proposed CDEC has been compared with the adversarial autoencoder,constrained 1-spectral clustering and autoencoder t k-means was applied to the known MNIST,Reuters-10k and USPS datasets,and their performance were assessed in terms of clustering *** results confirmed the statistical superiority of CDEC in terms of clustering accuracy to the ***-First of all,this is the first DNN-constrained clustering that

关键词： Deep neural networks Clustering Constrained clustering Big data denoising autoencoder Kullback-Leibler divergence Constrained deep embedded clustering(CDEC)

来源：评论

学校读者我要写书评

暂无评论

SNR-Aware Convolutional Neural Network Modeling for Speech Enhancement 17

SNR-Aware Convolutional Neural Network Modeling for Speech E...

引用

17th Annual Conference of the International-Speech-Communication-Association (INTERSPEECH 2016)

作者： Fu, Szu-Wei Tsao, Yu Lu, Xugang Acad Sinica Res Ctr Informat Technol Innovat Taipei Taiwan Natl Taiwan Univ Dept Comp Sci & Informat Engn Taipei Taiwan Natl Inst Informat & Commun Technol Kyoto Japan

ISBN: (纸本)9781510833135

This paper proposes a signal-to-noise-ratio (SNR) aware convolutional neural network (CNN) model for speech enhancement (SE). Because the CNN model can deal with local temporal-spectral structures of speech signals, it can effectively disentangle the speech and noise signals given the noisy speech signals. In order to enhance the generalization capability and accuracy, we propose two SNR-aware algorithms for CNN modeling. The first algorithm employs a multi -task learning (MTL) framework, in which restoring clean speech and estimating SNR level are formulated as the main and the secondary tasks, respectively, given the noisy speech input. The second algorithm is an SNR adaptive denoising, in which the SNR level is explicitly predicted in the first step, and then an SNR-dependent CNN model is selected for denoising. Experiments were carried out to test the two SNR-aware algorithms for CNN modeling. Results demonstrate that CNN with the two proposed SNR-aware algorithms outperform the deep neural network counterpart in terms of standardized objective evaluations when using the same number of layers and nodes. Moreover, the SNR-aware algorithms can improve the de noising performance with unseen SNR levels, suggesting their promising generalization capability for real-world applications.

关键词： speech enhancement convolutional neural network denoising autoencoder multi-task learning

来源：评论

学校读者我要写书评

暂无评论

Denoised Labels for Financial Time Series Data via Self-Supervised Learning 22

Denoised Labels for Financial Time Series Data via Self-Supe...

引用

3rd ACM International Conference on AI in Finance (ICAIF)

作者： Ma, Yanqing Ventre, Carmine Polukarov, Maria Kings Coll London London England

ISBN: (纸本)9781450393768

The introduction of electronic trading platforms effectively changed the organisation of traditional systemic trading from quote-driven markets into order-driven markets. Its convenience led to an exponentially increasing amount of financial data, which is however hard to use for the prediction of future prices, due to the low signal-to-noise ratio and the non-stationarity of financial time series. Simpler classification tasks - where the goal is to predict the directions of future price movement via supervised learning algorithms need sufficiently reliable labels to generalise well. Labelling financial data is however less well defined than in other domains: did the price go up because of noise or a signal? The existing labelling methods have limited countermeasures against the noise, as well as limited effects in improving learning algorithms. This work takes inspiration from image classification in trading [6] and the success of self-supervised learning in computer vision (e.g., [16]). We investigate the idea of applying these techniques to financial time series to reduce the noise exposure and hence generate correct labels. We look at label generation as the pretext task of a self-supervised learning approach and compare the naive (and noisy) labels, commonly used in the literature, with the labels generated by a denoising autoencoder for the same downstream classification task. Our results demonstrate that these denoised labels improve the performances of the downstream learning algorithm, for both small and large datasets, while preserving the market trends. These findings suggest that with our proposed techniques, self-supervised learning constitutes a powerful framework for generating "better" financial labels that are useful for studying the underlying patterns of the market.

关键词： Self-Supervised Learning denoising autoencoder Financial Labels Directional-Change Events

来源：评论

学校读者我要写书评

暂无评论

An Agreement Based Dynamic Routing Method for Fault Diagnosis in Power Network with Enhanced Noise Immunity 31

An Agreement Based Dynamic Routing Method for Fault Diagnosi...

引用

31st Australasian Universities Power Engineering Conference (AUPEC)

作者： Fahim, Shahriar Rahman Muyeen, S. M. Sarker, Yeahia Sarker, Subrata K. Das, Sajal K. Amer Int Univ Bangladesh Dhaka 1229 Bangladesh Qatar Univ Doha 2713 Qatar Rajshahi Univ Engn & Technol Rajshahi 6204 Bangladesh

ISBN: (纸本)9781665434515

The stable operation of a power system often depends on inscribing the faults that may arise when transmitting and distributing electrical power. Characterizing these faults is necessary to analyze the post-fault oscillography of the power lines. The power lines are prone to be affected by noises. The noises are responsible to introduce uncertainty in operating conditions. The variation in operating conditions leads to an unbalanced system. The diagnosis of faults is essential to ensure the secured operation of a power network. This paper introduces a unified unsupervised learning framework for short circuit fault analysis of a power transmission line. The proposed approach works with a small number of data set and reduces the computational cost. It uses a capsule network that investigates the low-level fault-oriented features. To guarantee the robustness of the proposed framework against noises a stacked denoising-autoencoder is integrated and modeled. The performance of the proposed model is measured and compared with some of the techniques available in the literature in terms of noise. The test with field data for three types of fault classification results in an accuracy of 9 ms for fault triggering.

关键词： Faults transmission line Capsule network time series imaging vector denoising autoencoder

来源：评论

学校读者我要写书评

暂无评论

Stochastic learning with Back Propagation

Stochastic learning with Back Propagation

引用

IEEE International Symposium on Circuits and Systems (IEEE ISCAS)

作者： Kim, Guhyun Hwang, Cheol Seong Jeong, Doo Seok Korea Inst Sci & Technol Ctr Elect Mat Seoul South Korea Seoul Natl Univ Sch Mat Sci & Engn Seoul South Korea Hanyang Univ Div Mat Sci & Engn Seoul South Korea

ISBN: (纸本)9781728103976

Despite of remarkable progress on deep learning, its hardware implementation beyond deep learning acceleration is still behind the software deep learning due in part to lack of hardware-compatible learning algorithm. In this paper, a learning method called the stochastic learning with backpropagation (SLBP) algorithm was proposed. The network of concern consists of ternary synaptic weight, favorable to be implemented in a resistance-based crossbar array. Every training epoch, the SLBP algorithm evaluates weight update probability at which the corresponding weight is updated in a stochastic manner. The algorithm was used to train a denoising autoencoder, which identified the successful reduction in noise (increase in peak signal-to-noise ratio by approximately 68%). Notably, the SLBP algorithm achieves an 86% reduction in memory usage compared with a real-valued autoencoder trained using a backpropagation algorithm.

关键词： Stochastic learning with backpropagation ternary weight crossbar array denoising autoencoder

来源：评论

学校读者我要写书评

暂无评论

Single channel speech enhancement using convolutional neural network

Single channel speech enhancement using convolutional neural...

引用

IEEE International Workshop of Electronics Control Measurement Signals and their Application to Mechatronics (ECMSM)

作者： Kounovsky, Tomas Malek, Jiri Tech Univ Liberec Fac Mechatron Inst Informat Technol & Elect Liberec Czech Republic

ISBN: (纸本)9781509055821

Neural networks can be used to identify and remove noise from noisy speech spectrum (denoisisng autoencoders, DAEs). The DAEs are typically implemented using the fully-connected feed-forward topology. Usually one of the following possibilities is used as DA target: 1) Ideal frequency ratio mask, which is applied to noisy spectrum to estimate the clean speech spectrum (masking) or 2) Clean speech spectrum directly (mapping). Recent research in the area of automatic speech recognition shows that convolutional neural networks are very promising in speech modeling. In this paper we thus suggest, construct the DAs with the convolutional topology. We investigate the suitability of both above described target types and compare the results with the fully connected DAs. Our experiments indicate that mapping-based convolutional networks estimating log-power spectra achieve significant improvement over all competing topologies and target types. Performance gains of 8% in PESQ scores over the ideal ratio masking are observed. We also investigate the ability of DAEs to enhance speech of unseen language based on the language diversity of the training set. Our experiments suggest that training DAEs with language-diverse training sets does not yield any significant benefit for the task of speech enhancement.

关键词： denoising autoencoder Single channel speech enhancement Convolutional neural network Multilingual training

来源：评论

学校读者我要写书评

暂无评论

Cleaning Noisy ECG based on the Signal Quality with Single and Multiple Hidden Layer autoencoder 2

Cleaning Noisy ECG based on the Signal Quality with Single a...

引用

2nd International Conference on Intelligent Cybernetics Technology and Applications (ICICyTA) - Cybernetic Technology for Better Future of Human Life

作者： Marwan, Bagas Samann, Fars Schanze, Thomas Univ Appl Sci THM Fac Life Sci Engn LSE Inst Biomed Engn IBMT Giessen Germany Univ Duhok Dept Biomed Engn Duhok Kurdistan Regio Iraq

ISBN: (纸本)9798350399134

Bio-signal recordings, e.g., electrocardiogram (ECG), often interfere with noise from different sources, which could affect the morphology of essential features, such as P-wave and T-wave. These features are superimposed with noise and make it difficult to interpret them. Here, we target to improve the quality of the ECG features with unsupervised neural networks, such as denoising autoencoder (DAE). In this work, the training and testing performance of single hidden layer and multiple hidden layers denoising autoencoders (SHL- and MHL-DAE) is evaluated based on the signal quality and the restoration of the ECG morphology waveforms by using signal-to-noise ratio improvement and the characteristic morphology of the ECG waves. The DAE was trained and tested with the superposition of simulated ground truth ECG and recorded physical noises, such as baseline wandering noise, electrode movement noise, muscle artifacts noise, a mixture of noises, and gaussian white noise. The effects of using different activation functions, the number of neurons, the number of the dataset, and dataset pre-processing are presented in this work. In the case of jittered-segment datasets, the MHL-DAE performs better than the SHL-DAE. However, the selection of the essential activation function and set of hyperparameters, such as the number of hidden neurons, must be done carefully.

关键词： denoising autoencoder filtering activation function hidden neurons jittering physical noise

来源：评论

学校读者我要写书评

暂无评论

A Deep Learning Technique for Electricity Price Forecasting in Consideration of Spikes

A Deep Learning Technique for Electricity Price Forecasting ...

引用

IEEE Region 10 Conference (TENCON)

作者： Yamada, Kodai Mori, Hiroyuki Meiji Univ Dept Network Design Tokyo 1648525 Japan

ISBN: (纸本)9781665495325

This paper presents a Deep Neural Network (DNN) method for electricity price forecasting in power markets. They are inclined to generate spikes that are dozens to hundred times as large as the normal prices so that the prediction is hard to handle. This paper focuses attention on the prediction of spikes to suppress the forecasting errors. This paper deals with the pretraining technique of autoencoder (AE) in Deep Learning. To enhance the performance of AE, this paper presents a denoising-autoencoder (DAE)-based method that consists of DAE and Multilayer Perceptron (MLP) of ANN with the clustering technique. DAE is an extension of AE in a sense that noisy learning data is used with random numbers. The use of clustering enhances the model accuracy due to data similarity. The effectiveness of the proposed method is tested for data of New England ISO, USA.

关键词： electricity price forecasting deep neural network autoencoder denoising autoencoder LMP power markets spikes

来源：评论

学校读者我要写书评

暂无评论

autoencoder-based Semi-Supervised Curriculum Learning For Out-of-domain Speaker Verification 20

Autoencoder-based Semi-Supervised Curriculum Learning For Ou...

引用

Interspeech Conference

作者： Zheng, Siqi Liu, Gang Suo, Hongbin Lei, Yun Alibaba Grp Machine Intelligence Technol Hangzhou Peoples R China

This study aims to improve the performance of speaker verification system when no labeled out-of-domain data is available. An autoencoder-based semi-supervised curriculum learning scheme is proposed to automatically cluster unlabeled data and iteratively update the corpus during training. This new training scheme allows us to (1) progressively expand the size of training corpus by utilizing unlabeled data and correcting previous labels at run-time;and (2) improve robustness when generalizing to multiple conditions, such as out-of-domain and text-independent speaker verification tasks. It is also discovered that a denoising autoencoder can significantly enhance the clustering accuracy when it is trained on carefully-selected subset of speakers. Our experimental results show a relative reduction of 30% - 50% in EER compared to the baseline.

关键词： Speaker Verification Semi-Supervised Learning Curriculum Learning denoising autoencoder SSCL

来源：评论

学校读者我要写书评

暂无评论

Data-driven Prior for Pharmaceutical Snapshot Spectral Imaging 25

Data-driven Prior for Pharmaceutical Snapshot Spectral Imagi...

引用

25th IEEE International Conference on Computational Science and Engineering (CSE)

作者： Su, Xuesan Mao, Jianxu Wang, Yaonan Chen, Yurong Zhang, Hui Hunan Univ NERC RVC Changsha Peoples R China

ISBN: (纸本)9798350396331

This paper proposes a new method for pharmaceutical hyperspectral compressive imaging and has a significant improvement for the quality of reconstruction. It's known that coded aperture snapshot spectral imager(CASSI) overcomes the limitation of hyperspectral image acquisition. However, the spatial and spectral information is coded and overlapped which make it difficult to reconstruct the original images. The reconstruction is an inverse mathematical problem which is barely solved precisely especially in complex imaging scenes such as irregular pharmaceutical product imaging. Thus, we consider the real pharmaceutical imaging demands and propose a novel image restoration method with the data-driven prior. Our method is based on the generalized alternating projection(GAP) framework and propose a novel denoising part to solve the problem of detail texture feature extraction with the dense block module employed. Our method is tested on real pharmaceutical hyperspectral data and achieve higher performance compared with state of the art methods.

关键词： computational imaging compressive sensing denoising autoencoder dense block data-driven prior

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：