检索结果-内蒙古大学图书馆

Mode shape prediction based on convolutional neural network and autoencoder

STRUCTURES 2022年 40卷 127-137页

作者： Hu, Kejian Wu, Xiaoguang Changan Univ Highway Sch Xian 71000 Peoples R China

Mode shape is a dynamic characteristic that plays an important role in civil engineering. In this paper, an approach to predict the mode shape of a bridge is proposed using a convolutional neural network (CNN) and an autoencoder. First, a large mode shape database of a bridge is established by the finite element method for training networks. Second, a mode shape tensor is formed based on the mode-shape results. Then, an autoencoder is trained to encode the tensor to a three-dimensional latent-space representation and restore it from the representations. The CNN can output the representation directly rather than the mode shape to reduce the training difficulty and improve the accuracy. The CNN takes 18 bridge design parameters and an original shape tensor, which is constructed based on 16 geometric parameters. An evaluation of the test set shows that the approach can predict the first three order mode shapes well, with the accuracy of 0.92, 0.83 and 0.79, while performs poorly in the fourth and fifth orders, with the accuracy of 0.68 and 0.63. In addition, the spatial distribution of the latent space representation is explored. The necessity of an autoencoder and the original shape tensor is demonstrated.

关键词： Mode shape Convolutional neural network autoencoder Bridge engineering

来源：评论

学校读者我要写书评

暂无评论

Semi-Federated Learning Based on autoencoder for Non-Intrusive Load Monitoring 23

Semi-Federated Learning Based on Autoencoder for Non-Intrusi...

引用

23rd IEEE International Conference on Communication Technology, ICCT 2023

作者： Bao, Chongyu Zhao, Haitao Li, Ruize Xia, Wenchao Nanjing University of Posts and Telecommunications Nanjing China

ISBN: (纸本)9798350325959

To addresses the challenges of data scarcity and weak computational capabilities of edge devices in the practical application of non-intrusive load monitoring (NILM) of power systems, we propose a semi-federated learning approach using an autoencoder network based on Mel-scale frequency cepstral coefficients (MFCC), referred to as semi-FedNILM. We introduce a MFCC-based autoencoder network to ensure the completeness of hidden layer feature information and incorporate a semi-federated learning framework into the network structure, such that only the training process of the autoencoder is performed on the clients. The encoded and encrypted data is then sent to the cloud server for training the global NILM model, significantly reducing the computational burden on edge clients while preserving user privacy. Experimental results demonstrate that the proposed Semi-FedNILM algorithm achieves better performance on homogenous and heterogeneous datasets of a practical task. © 2023 IEEE.

关键词： autoencoder Federated learning Non-intrusive load monitoring Privacy preservation

来源：评论

学校读者我要写书评

暂无评论

Identification of Stealthy Hardware Trojans through On-Chip Temperature Sensing and an autoencoder-Based Machine Learning Algorithm 66

Identification of Stealthy Hardware Trojans through On-Chip ...

引用

2023 IEEE 66th International Midwest Symposium on Circuits and Systems, MWSCAS 2023

作者： Gourousis, Thomas Zhang, Ziyue Yan, Mengting Zhang, Millin Mittal, Ankit Shrivastava, Aatmesh Restuccia, Francesco Fei, Yunsi Onabajo, Marvin Northeastern University Dept. of Electrical and Computer Engineering BostonMA United States

ISBN: (纸本)9798350302103

This paper presents an anomaly detection approach with non-invasive on-chip temperature sensing for hardware Trojan detection, which is coupled with a proposed anomaly detection technique using an autoencoder-based machine learning (ML) algorithm. In a case study, the developed algorithm identifies the Hardware Trojan with over 90% accuracy for a Trojan power consumption that is as low as 2.5% of the circuit under test. Besides identifying the Trojan, the algorithm can also provide information about the location of the Trojan on the chip. The proposed on-chip anomaly detection approach with machine learning is under development as a solution for enhanced hardware security in modern electronic systems, particularly for Internet of Things applications. © 2023 IEEE.

关键词： autoencoder Hardware Trojan detection machine learning on-chip temperature sensing trusted electronics

来源：评论

学校读者我要写书评

暂无评论

Object Attention autoencoder 9

Object Attention Autoencoder

引用

9th IEEE Smart World Congress, SWC 2023

作者： Bi, Yanqing Luo, Yu Wang, Zhenyu National University of Defense Technology College of Computer Changsha China

ISBN: (纸本)9798350319804

autoencoder (AE) networks are utilized in novelty detection, classification, and deep clustering tasks to learn feature representation. While AEs have demonstrated promising performance in various applications, we observe that existing AE algorithms do not well in focusing more on the reconstruction of the foreground. To address this issue, this paper proposes an Object Attention autoencoder (OAE) algorithm. Specifically, we use the background-removed and original image to guide each other, ensuring that the model is more focused on the reconstruction of foreground information while producing consistent latent layer features for the two types image. Subsequently, a distance-oriented loss is directly built on the latent layer to bring the features of the background-removed image and the original image close to each other. In short, we simultaneously minimize the reconstruction loss and the distance loss. The resultant optimization problem can be effectively solved by means of stochastic gradient descent and backpropagation. Experiments on the MS COCO dataset empirically validate that our method pays more attention to foreground information while producing a robust latent layer feature. © 2023 IEEE.

关键词： autoencoder feature representation object attention

来源：评论

学校读者我要写书评

暂无评论

Pulsar candidate recognition based on autoencoder and self-normalizing neural networks 11

Pulsar candidate recognition based on autoencoder and self-n...

引用

11th Joint International Information Technology and Artificial Intelligence Conference, ITAIC 2023

作者： Wang, Zhihao Fu, Peng Guo, Meng Shandong Academy of Sciences Shandong Jinan250013 China Shandong Provincial Key Laboratory of Computer Networks Shandong Fundamental Research Center for Computer Science Shandong Jinan250013 China

ISBN: (纸本)9798350333664

The identification of pulsar candidates is a crucial step in radio astronomy research. With the continuous improvement of modern radio telescope equipment and the increasing scale of pulsar sky survey, a pulsar survey will produce a large number of pulsar candidates. Identifying pulse signals from a large number of pulsar candidates is a prerequisite for discovering new pulsar. In the face of massive pulsar candidate data, it is a huge workload to identify pulsar candidates manually only by experts in relevant fields. In order to improve the efficiency of identifying pulsar candidates, a stacked autoencoder and self-normalizing neural network based pulsar candidate recognition framework AESNN is proposed, which does not need to consider the imbalance of the number of positive and negative samples in the pulsar data set. The pulsar candidate contains many features, but not all of them are important. The stacked autoencoder is used to extract the feature representation of the data, perform unsupervised dimensionality reduction, and input these features into the self-normalizing neural network for training. AESNN introduces self-normalizing properties by introducing 'scaled exponential linear units', and experiments show that AESNN has a high F1 score, precision, and recall, and does not require the use of complex data augmentation methods, making it more advantageous compared to other algorithms. © 2023 IEEE.

关键词： autoencoder Deep learning pulsar self-normalizing

来源：评论

学校读者我要写书评

暂无评论

Abnormal condition monitoring method of cement pre-decomposition process based on PCA and autoencoder network 38

Abnormal condition monitoring method of cement pre-decomposi...

引用

38th Youth Academic Annual Conference of Chinese Association of Automation, YAC 2023

作者： Dong, Huijun Ma, Zongxian Lu, Shizeng Yu, Hongliang University of Jinan School of Electrical Engineering Jinan250022 China

ISBN: (纸本)9798350303636

The accurate monitoring of abnormal production conditions in cement process is the basis of intelligent control, which is of great significance to improve the intelligent level of cement production. In this paper, an abnormal condition monitoring method of cement pre-decomposition process based on PCA and autoencoder network is proposed. In order to improve the accuracy of anomaly detection, this paper integrates the PCA and autoencoder algorithms. Firstly, the anomaly detection models based on PCA and autoencoder are constructed. On this basis, anomaly monitoring thresholds based on PCA and autoencoder are determined. For the samples to be tested, the PCA-based and autoencoder-based reconstruction errors are obtained and compared with the reconstruction error thresholds. If any of the judgments are abnormal, the test sample is in an abnormal state. Finally, taking the kiln tail crusting condition in the cement pre-decomposition link as an example, the feasibility of the proposed method is verified. The identification results shows that for the 440 test samples, only 4 abnormal points are not identified for the abnormal working conditions of kiln tail crusting. This paper provided a feasible method for abnormal condition monitoring of cement pre-decomposition process. © 2023 IEEE.

关键词： abnormal condition monitoring autoencoder cement pre-decomposition process PCA

来源：评论

学校读者我要写书评

暂无评论

A Supervised autoencoder for Human Activity Recognition with Inertial Sensors

A Supervised Autoencoder for Human Activity Recognition with...

引用

2023 IEEE International Conference on Big Data, BigData 2023

作者： An, Jaehyuk Kwon, Younghoon Cho, Yoon-Sik Chung-Ang University Korea Republic of University of Washington United States

ISBN: (纸本)9798350324457

Human Activity Recognition (HAR) with inertial sensors is one of the most active research fields. Various machine learning algorithms have been proposed in HAR for classifying human activities. However, these methods heavily rely on the quality of hand-crafted features, requiring extensive feature engineering. Recent deep learning approaches have tried to perform training in an end-to-end manner. We propose a new learning scheme based on Supervised autoencoder with Self-Attention (SAE-SA). Our main idea is two-fold: (1) Through the dimensional reduction in supervised autoencoder, our model is robust to noisy input signals sensor data (2) We incorporate the self-attention mechanism in the classifier layer of supervised autoencoder, which focuses more to the relevant signals, and can learn the features without any knowledge of the aggregated signal data. We evaluate SAE-SA on benchmark datasets: WISDM v2.0, and PAMAP2. We achieve accuracy of 95.76%, and 97.60% respectively, which is the state-of-the-art results in HAR. © 2023 IEEE.

关键词： autoencoder Deep learning Human Activity Recognition Self-Attention Supervised learning

来源：评论

学校读者我要写书评

暂无评论

A CGAN-aided autoencoder Supporting Joint Geometric Probabilistic Shaping for Optical Fiber Communication System

A CGAN-aided Autoencoder Supporting Joint Geometric Probabil...

引用

2023 Asia Communications and Photonics Conference/2023 International Photonics and Optoelectronics Meetings, ACP/POEM 2023

作者： Li, Yuzhe Chang, Huan Zhang, Qi Xin, Xiangjun Gao, Ran Tian, Feng Tian, Qinghua Wang, Fu Li, Zhipei Beijing University of Posts and Telecommunications Beijing Key Laboratory of Space-Ground Interconnection and Convergence Beijing China School of Information and Electronics Beijing Institute of Technology Beijing China

ISBN: (纸本)9798350312614

An autoencoder supporting joint geometric and probabilistic shaping is proposed that can achieve global optimization of optical fiber communication systems with aid of conditional generative adversarial network. Result shows that bit error rate is reduced by 12% compared to a probabilistic shaping-only signal with the same entropy. © 2023 IEEE.

关键词： autoencoder channel modeling geometric shaping probabilistic shaping

来源：评论

学校读者我要写书评

暂无评论

Improved Knowledge Transfer based on Stacked Denoising autoencoder and Graph Convolutional Network for Fault Diagnosis of Rolling Bearings 14

Improved Knowledge Transfer based on Stacked Denoising Autoe...

引用

14th IEEE Global Reliability and Prognostics and Health Management Conference, PHM-Hangzhou 2023

作者： Yang, Tianyu Guo, Wei School of Mechanical and Electrical Engineering University of Electronic Science and Technology of China Chengdu611731 China

ISBN: (纸本)9798350301359

Intelligent fault diagnosis provides a useful tool for rotating machinery to ensure its safety and condition monitoring, which is a hot topic in prognostics and health management. Although the deep learning algorithms improve the clustering ability for data without labels, its learning efficiency and generalization ability are still limited. To solve this problem, an improved knowledge transfer method is presented based on a stacked denoising autoencoder and graph convolutional network (SDAE-GCN). Stacked structure in SDAE and structure information from GCN improve its learning ability and efficiency. Two open-source bearing datasets are used to evaluate the performance of the proposed SDAE-GCN. The experimental results demonstrate that it has higher accuracy and better generalization for bearing fault diagnosis when the training and test data sets have obvious differences, including different damage sizes/levels, rotation speeds, and working conditions. Noise sensitivity analysis and comparison further show its superiority. © 2023 IEEE.

关键词： autoencoder deep clustering fault diagnosis graph convolutional network rolling bearing transfer learning

来源：评论

学校读者我要写书评

暂无评论

Practical Application of Deep Modified autoencoder Technique to Electricity Price Forecasting

Practical Application of Deep Modified Autoencoder Technique...

引用

2023 IEEE International Conference on Energy Technologies for Future Grids, ETFG 2023

作者： Yamada, Kodai Mori, Hiroyuki Tepco Power Grid Inc. Tokyo General Branch Tokyo Japan Meiji University Dept. of Network Design Tokyo Japan

ISBN: (纸本)9781665471640

This paper proposes a method for electricity price forecasting (EPF) with Deep Modified autoencoder (DMAE). It is based on a deep model of the modified autoencoder (MAE) that improves the learning process by adding noise to the original data. EPF is hard to handle due to higher volatility in comparison with other time series in power systems. autoencoder (AE) plays a key role in reducing the dimensionality of input variables in Deep Neural Network (DNN), which is useful for constructing reasonable models with fewer parameters. In this paper, DMAE is proposed as a pretraining technique in Deep Neural Network (DNN). It may be regarded as an integration model that combines MAE with Stacked AE. The proposed method integrates DMAE with Multi-layer perceptron (MLP) to construct DNN. The effectiveness of the proposed method is tested for real data in Boston, MA, USA. © 2023 IEEE.

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

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：