检索结果-内蒙古大学图书馆

Confidence-Calibrated Human Activity Recognition

SENSORS 2021年第19期21卷 6566-6566页

作者： Roy, Debaditya Girdzijauskas, Sarunas Socolovschi, Serghei KTH Royal Inst Technol Sch Elect Engn & Comp Sci EECS S-11428 Stockholm Sweden

Wearable sensors are widely used in activity recognition (AR) tasks with broad applicability in health and well-being, sports, geriatric care, etc. Deep learning (DL) has been at the forefront of progress in activity classification with wearable sensors. However, most state-of-the-art DL models used for AR are trained to discriminate different activity classes at high accuracy, not considering the confidence calibration of predictive output of those models. This results in probabilistic estimates that might not capture the true likelihood and is thus unreliable. In practice, it tends to produce overconfident estimates. In this paper, the problem is addressed by proposing deep time ensembles, a novel ensembling method capable of producing calibrated confidence estimates from neural network architectures. In particular, the method trains an ensemble of network models with temporal sequences extracted by varying the window size over the input time series and averaging the predictive output. The method is evaluated on four different benchmark HAR datasets and three different neural network architectures. Across all the datasets and architectures, our method shows an improvement in calibration by reducing the expected calibration error (ECE)by at least 40%, thereby providing superior likelihood estimates. In addition to providing reliable predictions our method also outperforms the state-of-the-art classification results in the WISDM, UCI HAR, and PAMAP2 datasets and performs as good as the state-of-the-art in the Skoda dataset.

关键词： wearable sensors human activity recognition deep learning confidence calibration time-series classification signal processing model reliability training algorithm

来源：评论

学校读者我要写书评

暂无评论

Momentum Acceleration of Quasi-Newton training for Neural Networks 16th

Momentum Acceleration of Quasi-Newton Training for Neural Ne...

引用

16th Pacific Rim International Conference on Artificial Intelligence (PRICAI)

作者： Mahboubi, Shahrzad Indrapriyadarsini, S. Ninomiya, Hiroshi Asai, Hideki Shonan Inst Technol 1-1-25 Tsujido Nishikaigan Fujisawa Kanagawa 2518511 Japan Shizuoka Univ Naka Ku 3-5-1 Johoku Hamamatsu Shizuoka 4328011 Japan

ISBN: (纸本)9783030299118;9783030299101

This paper describes a novel acceleration technique of quasi-Newton method (QN) using momentum terms for training in neural networks. Recently, Nesterov's accelerated quasi-Newton method (NAQ) has shown that the momentum term is effective in reducing the number of iterations and in accelerating its convergence speed. However, the gradients had to calculate two times during one iteration in the NAQ training. This increased the computation time of a training loop compared with the conventional QN. In this research, an improvement to NAQ is done by approximating the Nesterov's accelerated gradient used in NAQ as a linear combination of the current and previous gradients. Then the gradient is calculated only once per iteration same as QN. The performance of the proposed algorithm is evaluated through computer simulations on a benchmark problem of the function modeling and real-world problems of the microwave circuit modeling. The results show the significant acceleration in the computation time compared with conventional training algorithms.

关键词： Neural networks training algorithm Quasi-Newton method Nesterov's accelerated quasi-Newton method Momentum terms

来源：评论

学校读者我要写书评

暂无评论

training GENERATIVE ADVERSARIAL NETWORKS WITH WEIGHTS 27

TRAINING GENERATIVE ADVERSARIAL NETWORKS WITH WEIGHTS

引用

27th European Signal Processing Conference (EUSIPCO)

作者： Pantazis, Yannis Paul, Dipjyoti Fasoulakis, Michail Stylianou, Yannis FORTH Inst Appl & Computat Math Iraklion Greece Univ Crete Dept Comp Sci Iraklion Greece FORTH Inst Comp Sci Iraklion Greece

ISBN: (纸本)9789082797039

The impressive success of Generative Adversarial Networks (GANs) is often overshadowed by the difficulties in their training Despite the continuous efforts and improvements, there are still open issues regarding their convergence properties. In this paper, we propose a simple training variation where suitable weights are defined and assist the training of the Generator. We provide theoretical arguments which indicate that the proposed algorithm is better than the baseline algorithm in the sense of creating a stronger Generator at each iteration. Performance results showed that the new algorithm is more accurate and converges faster in both synthetic and image datasets resulting in improvements ranging between 5% and 50%.

关键词： Generative adversarial networks multiplicative weight update method training algorithm

来源：评论

学校读者我要写书评

暂无评论

Freeze-drying behaviour prediction of button mushrooms using artificial neural network and comparison with semi-empirical models

引用

NEURAL COMPUTING & APPLICATIONS 2019年第11期31卷 7257-7268页

作者： Tarafdar, Ayon Shahi, Navin Chandra Singh, Anupama Natl Inst Food Technol Entrepreneurship & Managem Dept Food Engn Sonipat 131028 Haryana India GB Pant Univ Agr & Technol Dept Postharvest Proc & Food Engn Pantnagar 263145 Uttar Pradesh India

The application of artificial neural networks (ANN) in the freeze-drying of button mushrooms has been investigated. Networks with a single hidden layer, different training algorithms and complexity in terms of the number of neurons were evaluated for identifying the best ANN infrastructure. Moisture content, moisture ratio and drying rate were taken as output drying parameters for which ANN models provided an overall correlation coefficient (R) of 0.994, 0.991 and 0.992, respectively. The predictive efficiency of ANN was compared to semi-empirical models. Coefficients for semi-empirical models of moisture ratio were determined. Logarithm model gave the best fit (R-2 = 0.985) for moisture ratio prediction but with larger mean square error and lower correlation than ANN model. The study highlights that ANN models with low complexity can be developed to precisely predict drying behaviour of biological materials while providing comparable and even superior results to that obtained from available semi-empirical drying models.

关键词： Artificial neural network training algorithm Freeze-drying Button mushroom

来源：评论

学校读者我要写书评

暂无评论

Vocoder-free text-to-speech synthesis incorporating generative adversarial networks using low-/multi-frequency STFT amplitude spectra

引用

COMPUTER SPEECH AND LANGUAGE 2019年 58卷 347-363页

作者： Saito, Yuki Takamichi, Shinnosuke Saruwatari, Hiroshi Univ Tokyo Grad Sch Informat Sci & Technol Bunkyo Ku 7-3-1 Hongo Tokyo 1138656 Japan

This paper proposes novel training algorithms for vocoder-free text-to-speech (TTS) synthesis based on generative adversarial networks (GANs) that compensate for short-term Fourier transform (STFT) amplitude spectra in low/multi frequency resolution. Vocoder-free TTS using STFT amplitude spectra can avoid degradation of synthetic speech quality caused by the vocoder-based parameterization used in conventional TTS. Our previous work for the vocoder-based TTS proposed a method for incorporating the GAN-based distribution compensation into acoustic model training to improve synthetic speech quality. This paper extends the algorithm to the vocoder-free TTS and propose a GAN-based training algorithm using low-frequency-resolution amplitude spectra to overcome the difficulty in modeling complicated distribution of the high-dimensional spectra. In the proposed algorithm, amplitude spectra are transformed into low-frequency-resolution amplitude spectra by applying an average pooling function along with a frequency axis;then the GAN-based distribution compensation is performed in the low-frequency-resolution domain. Because the low-frequency-resolution amplitude spectra approximately emulate filter banks, the proposed algorithm is expected to improve synthetic speech quality by reducing differences in spectral envelopes of natural and synthetic speech. Furthermore, various frequency scales that are related to human speech perception (e.g., mel and inverse mel frequency scales) can be introduced to the proposed training algorithm by applying an frequency warping function to amplitude spectra. This paper also proposes a GAN-based training algorithm using multi-frequency-resolution amplitude spectra that uses both low- and original-frequency-resolution amplitude spectra to reduce the differences in not only spectral envelopes but also fine structures. Experimental results demonstrate that (1) GANs using low-frequency-resolution amplitude spectra improve speech quality and work

关键词： Vocoder-free text-to-speech training algorithm STFT amplitude spectra Generative adversarial networks Frequency resolution Frequency warping

来源：评论

学校读者我要写书评

暂无评论

A fuzzy process neural network model and its application in process signal classification

引用

NEUROCOMPUTING 2019年 335卷 1-8页

作者： Xu, Shaohua Liu, Kun Li, Xuegui Shandong Univ Sci & Technol Coll Comp Sci & Engn Qingdao Peoples R China

Aiming at process signal classification combined with fuzzy decision rules, a fuzzy process neural network (FPNN) is proposed in this paper. The FPNN is structured with a process signal input layer, a fuzzy process neuron (FPN) hidden layer, a signal pattern layer, and a fuzzy decision output layer. The spatio-temporal aggregation in the FPN is taken as a generalized inner product operation, which can measure the fuzzy similarity of the distributed features among process signals and FPN uses an exponential fuzzy membership function as activation function. The FPNN selectively sums the output of the FPN hidden layer to the pattern layer according to the category of input signal. By linking a Takagi-Sugeno fuzzy classifier behind the pattern layer, direct classification of the process signals is achieved. The FPNN can improve the deficiencies of existing time-varying signal classification methods, such as the complete training data sets are required, and the information processing processes and algorithms are complicated. In this paper, the theoretical properties of the FPNN are analyzed, and the comprehensive learning algorithm for FPNNs is given. The discrimination of reservoir water flooding condition based on multi-channel well logging process signals was used as an example for experimental analysis;the results verify the validity of the model and algorithm. (C) 2019 Elsevier B.V. All rights reserved.

关键词： Process signal classification Fuzzy decision Fuzzy process neural network training algorithm Small sample modeling

来源：评论

学校读者我要写书评

暂无评论

A Probabilistic Process Neural Network and Its Application in ECG Classification

引用

IEEE ACCESS 2019年 7卷 50431-50439页

作者： Feng, Naidan Xu, Shaohua Liang, Yongquan Liu, Kun Shandong Univ Sci & Technol Coll Comp Sci & Engn Qingdao 266590 Shandong Peoples R China

A novel probabilistic process neural network (PPNN) model is proposed for the multi-channel time-varying signal classification problems with ambiguity and randomness distribution characteristics. This model was constructed from an input time-varying signal layer, a probabilistic process neuron (PPN) hidden layer, a pattern layer, and a Softmax classifier. The number of nodes in the input layer is the same as the number of time-varying signal input channels, which can realize the overall input of the time-varying signal. The PPN hidden layer performs the spatio-temporal aggregation operations for time-varying input signals and probabilistic outputs. The connection weight functions from the input layer to the hidden layer are represented by the typical samples or cluster center functions in different pattern subsets of the sample set, which the prior knowledge of signal categories is implicitly expressed by the morphological distribution features and combinational relations. A pattern layer selectively summed to the output of the PPN hidden layer using the categorical attributes of the connection weight function vector. And the Softmax classifier implements the probabilistic classification of time-varying signals. PPNN has the advantages of fewer model parameters, suitable for modeling small samples and integrating prior knowledge of signal categories. This paper develops the specific learning algorithms which synthesize dynamic timewarping, C-means clustering, and BP algorithm. The 12-lead electrocardiogram (ECG) signals for heart disease diagnosis were used for classification testing. The experimental results from 12-lead ECG signal across ten types of disease classification verify the effectiveness of the model and the proposed algorithm.

关键词： Dynamic signal classification Bayesian decision rules probabilistic process neural network training algorithm ECG signal classification

来源：评论

学校读者我要写书评

暂无评论

Automatic sedimentary microfacies identification from logging curves based on deep process neural network

引用

CLUSTER COMPUTING-THE JOURNAL OF NETWORKS SOFTWARE TOOLS AND APPLICATIONS 2019年第5-Sup期22卷 12451-12457页

作者： Liu, Hui Xu, Shaohua Ge, Xinmin Liu, Jianyu Zahid, Muhammad Aleem Shandong Univ Sci & Technol Coll Min & Safety Engn Qingdao 266590 Shandong Peoples R China China Univ Petr East China Sch Geosci Qingdao 266580 Shandong Peoples R China Lasbela Univ Agr Water & Marine Sci Fac Marine Sci Uthal 90250 Balochistan Pakistan

An automatic sedimentary microfacies identification technique is developed based on the deep process neural network (DPNN), which consist several neurons and general non-time-varying neurons arranged in a certain topological structure. In this technique, the features of the shape and amplitude of logging curves are considered to form the category outputs. Combined with the deep learning theory, the diversity of the process features of logging curves and the complexity of combined features of multiple geophysical logging information are considered, and DPNN is created through the stacked superimposition of deep belief network and BP classifier. The technique maintains the structure and information relevance of process signal data and can characterize the distribution features of logging curves automatically, and classify the process signals directly. The theoretical nature and performance of the improved algorithm is tested and validated by some field examples.

关键词： Sedimentary microfacies identification Well logging data Deep learning Process neural network training algorithm

来源：评论

学校读者我要写书评

暂无评论

A radial basis probabilistic process neural network model and corresponding classification algorithm

引用

APPLIED INTELLIGENCE 2019年第6期49卷 2256-2265页

作者： Liu, Kun Xu, Shaohua Feng, Naidan Shandong Univ Sci & Technol Dept Comp Sci & Engn Qingdao 266590 Shandong Peoples R China

A radial basis probabilistic process neuron (RBPPN) and radial basis probabilistic process neural network (RBPPNN) model are proposed to fuse a priori knowledge for application to time-varying signal pattern classification. RBPPN inputs were multi-channel time-varying signals and a generalized inner product was used to perform spatio-temporal aggregation of input signals in the kernel. Typical signal samples from various pattern subsets in the sample set were used as kernel center functions, which use morphological distribution characteristics and combination relationships to implicitly express prior knowledge for the signal category. The exponential probability function was used as the activation function to achieve kernel transformation and RBPPN probability output. The RBPPNN is composed of process signal input layers, an RBPPN hidden layer, a pattern layer, and a Softmax classifier developed through stacking. Generalized inner product operations were used to conduct probability similarity measurements of distribution characteristics between process signals. The pattern layer selectivity summed inputs from the RBPPN hidden layer to the pattern layer according to the category of the kernel center function. Its outputs were then used as inputs in the Softmax classifier. The proposed RBPPNN information processing mechanism was extended to the time domain, and through learning time-varying signal training samples, achieved extraction, expression, and information association of time-varying signal characteristics, as well as direct classification. It can improve the deficiencies of existing neural networks, such as a complete large-scale training dataset is needed, and the information processing flow is complex. In this paper, the properties of the RBPPNN are analyzed and a specific learning algorithm is presented which synthesizes dynamic time warping, dynamic C-means clustering, and the mean square error algorithm. A series of 12-lead electrocardiogram (ECG) signals

关键词： Dynamic signal classification Radial basis probabilistic process neural network training algorithm ECG signal classification

来源：评论

学校读者我要写书评

暂无评论

Performance Analysis of Neural Network training algorithms and Support Vector Machine for Power Generation Forecast of Photovoltaic Panel

引用

IEEE LATIN AMERICA TRANSACTIONS 2017年第6期15卷 1091-1100页

作者： da Silva, T. V. Monteiro, R. V. A. Guimaraes, G. C. Moura, F. A. M. Albertini, M. R. M. C. Tamashiro, M. A. Univ Fed Uberlandia Dept Engn Eletr Uberlandia MG Brazil Univ Fed Uberlandia Programa Pos Grad Engn Eletr Uberlandia MG Brazil Univ Fed Uberlandia Uberlandia MG Brazil Univ Fed Triangulo Mineiro Dept Engn Eletr Uberaba Brazil Univ Fed Triangulo Mineiro Uberaba Brazil Inst Fed Tocantins IFTO Palmas Brazil

Current energy policies are encouraging the connection of power generation based on low-polluting technologies, mainly those using renewable sources to distribution networks. The photovoltaic (PV) systems have experienced a great growth around the word in last years. Hence, it becomes increasingly important to understand technical challenges, facing high penetration of PV systems at the grid, especially considering the effects of intermittence of this source on the power quality, reliability and stability of the electric distribution system. In another hand, the connections of distributed generators, by PV panels, changes voltage profile at low voltage power systems. This fact can affect the distribution networks on which they are attached causing overvoltage, undervoltage, frequency oscillations and changes in protection design. In order to predict these disturbs, because of this PV penetration, this article aims to analyze seven training algorithms used in artificial neural networks for temporal prediction of the generated active power and thus the state of the distribution network in which these microgenerators are connected and, then compare its best results with the Support Vector Machine (SVM) technique. As a result it was concluded that 3 algorithms are suitable for this type of analysis with the best performance among the seven analyzed was the Bayesian Regularization and that Artificial Neural Networks are more suitable for this problem than the SVM.

关键词： training algorithm Artificial Neural Network Support Vector Machine Photovoltaic penetration Microgeneration Power Quality

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：