检索结果-内蒙古大学图书馆

Road User Abnormal Trajectory Detection Using a deep autoencoder 1

13th International Symposium on Visual Computing (ISVC)

作者： Roy, Pankaj Raj Bilodeau, Guillaume-Alexandre Polytech Montreal LITIV Lab Dept Comp & Software Engn Montreal PQ Canada

ISBN: (数字)9783030038014

ISBN: (纸本)9783030038014;9783030038007

In this paper, we focus on the development of a method that detects abnormal trajectories of road users at traffic intersections. The main difficulty with this is the fact that there are very few abnormal data and the normal ones are insufficient for the training of any kinds of machine learning model. To tackle these problems, we proposed the solution of using a deep autoencoder network trained solely through augmented data considered as normal. By generating artificial abnormal trajectories, our method is tested on four different outdoor urban users scenes and performs better compared to some classical outlier detection methods.

关键词： deep autoencoder Unsupervised learning Data augmentation Abnormal trajectory detection

来源：评论

学校读者我要写书评

暂无评论

A Secure deep autoencoder-based 6G Channel Estimation to Detect/Mitigate Adversarial Attacks 5

A Secure Deep Autoencoder-based 6G Channel Estimation to Det...

引用

5th IEEE Global Power, Energy and Communication Conference (GPECOM)

作者： Oleiwi, Haider W. Mhawi, Doaa N. Al-Raweshidy, H. S. Brunel Univ London Dept Elect & Elect Engn Uxbridge UB8 3PH Middx England Middle Tech Univ Tech Inst Management Baghdad 10010 Iraq

ISBN: (纸本)9798350301984

Channel estimation (CE) is critical in wireless communications. However, it is vulnerable to adversarial attacks (AA) that are associated with the incorporated artificial intelligence (AI) functionality in 6G wireless communication systems/networks. The hazardous threat can compromise communications' confidentiality and integrity due to the expected infrastructure, features, and AI models of the 6G paradigm. This paper proposed a deep autoencoder (DAE)-based 6G CE model to detect and prevent AA. It was trained using a dataset generated from the MATLAB toolbox for AA and incorporated a secure transmission protocol. Simulations were conducted to evaluate the model's performance under different parameters (i.e., CE and DAE) with maximal epsilon values range (0.5-3.0). The results proved the model's sufficiency of accuracy and security to detect AA compared to existing CE techniques. The proposal provided a promising solution for a secure 6G DAE-based CE and showed robustness against AA. Additionally, it offered a feasible solution for the deep learning training data required and avoids overfitting. Overall, the proposed model provides a valuable contribution towards enhancing the security of 6G networks, and its performance should be further validated in real-world scenarios.

关键词： 6G Wireless Communication Networks Adversarial Attacks Artificial Intelligence Channel Estimation Cybersecurity deep autoencoder

来源：评论

学校读者我要写书评

暂无评论

deep autoencoder-like non-negative matrix factorization with graph regularized for link prediction in dynamic networks

引用

APPLIED SOFT COMPUTING 2023年 148卷

作者： Lv, Laishui Bardou, Dalal Liu, Yanqiu Hu, Peng China Jiliang Univ Coll Informat Engn Hangzhou 310018 Peoples R China Univ Abbes Laghrour Dept Comp Sci & Math ICOSI Lab Khenchela Algeria Anhui Normal Univ Sch Comp & Informat Wuhu 241002 Peoples R China

Link prediction in dynamic(temporal) networks refers to predicting future edges by analyzing the available network information. Among the existing temporal link prediction approaches, non-negative matrix factorization(NMF) is a kind of competitive algorithm and has attracted extensive attention. However, traditional NMF-based prediction methods are shallow methods and cannot fully mine the dynamic network, which may lead to a decrease in performance of algorithms. To overcome these shortcomings, inspired by deep autoencoder, we propose two novel deep autoencoder-like NMF with graph regularized prediction methods for dynamic networks. By fusing encoder component with deep structure into deep NMF model, our algorithms can sufficiently exploit the complex hierarchical information hidden in dynamic networks. To further extract the abundant information hidden in dynamic networks, graph regularization and PageRank are utilized to exploit the local and global topology information of each snapshot, respectively. By jointly optimizing them in deep autoencoder-like NMF model, our model is able to preserve the local and global information hidden in dynamic networks, simultaneously. Moreover, an effective alternating iterative method with convergence guarantee is developed for minimizing the established model. Finally, we test our proposed prediction methods on several synthetic and real world datasets to demonstrate that our approaches outperform the state-of-the-art prediction approaches.

关键词： Temporal networks Temporal link prediction deep NMF Graph regularization deep autoencoder

来源：评论

学校读者我要写书评

暂无评论

Predicting Earthquakes with Ionospheric Data: A Hybrid Approach Utilizing deep autoencoder and LSTM Networks 3

Predicting Earthquakes with Ionospheric Data: A Hybrid Appro...

引用

3rd International Conference on Computing and Machine Intelligence (ICMI)

作者： Abri, Rayan Abri, Sara Artuner, Harun Cetin, Salih Dept Artificial Intelligence Mavinci Ankara Turkiye Hacettepe Univ Dept Comp Engn Ankara Turkiye Mavinci Cambridge England

ISBN: (纸本)9798350372977;9798350372984

An integral part of the Earth's atmosphere is driven by the ionosphere. Solar flares induce ionosphere anomalies as a result of coronal mass ejection, seismic activity, and geomagnetic activity. Total Electron Content is the primary metric used to study the ionosphere's structure (TEC). GPS-derived TEC values are useful for examining how the ionospheric response to earthquakes is affected. In order to identify earthquakes, this article examines the relationships between TEC data and earthquakes. Our aim is to suggest a classification strategy for identifying earthquakes that occurred in earlier days. This research discusses the ionospheric variability during moderate and severe earthquake events of varied intensity for the years 2012-2019. deep autoencoders are used by the suggested model to extract features from TEC data. A Stacked LSTM model was constructed using the features gathered to forecast the earthquakes that occurred in the preceding days. For evaluation, the suggested hybrid model is compared with the Support Vector Machine (SVM) and Linear Discriminant Analysis (LDA) classifier models. According to the findings, the suggested hybrid model increases earthquake detection with an accuracy rate of roughly 0.84 and is a useful tool for identifying earthquakes based on prior days.

关键词： Total electron content Long short term memory deep autoencoder Support vector machine Linear Discriminant Analysis

来源：评论

学校读者我要写书评

暂无评论

Fault Detection of Elevator System Using Profile Extraction and deep autoencoder Feature Extraction for Acceleration and Magnetic Signals 23

Fault Detection of Elevator System Using Profile Extraction ...

引用

23rd International Conference on the Information Visualisation (IV) - Incorporating Biomedical Visualization, Learning Analytics and Geometric Modelling and Imaging

作者： Mishra, Krishna Mohan Krogerus, Tomi R. Huhtala, Kalevi J. Tampere Univ Unit Automat Technol & Mech Engn Tampere Finland

ISBN: (纸本)9781728128382

In this paper, we propose a new algorithm for data extraction from time series data, and furthermore automatic calculation of highly informative deep features to be used in fault detection. In data extraction elevator start and stop events are extracted from sensor data including both acceleration and magnetic signals. In addition, a generic deep autoencoder model is also developed for automated feature extraction from the extracted profiles. After this, extracted deep features are classified with random forest algorithm for fault detection. Sensor data are labelled as healthy and faulty based on the maintenance actions recorded. The remaining healthy data are used for validation of the model to prove its efficacy in terms of avoiding false positives. We have achieved nearly 100% accuracy in fault detection along with avoiding false positives based on new extracted deep features, which outperforms results using existing features. Existing features are also classified with random forest to compare results. Our developed algorithm provides better results due to the new deep features extracted from the dataset when compared to existing features. This research will help various predictive maintenance systems to detect false alarms, which will in turn reduce unnecessary visits of service technicians to installation sites.

关键词： Elevator System deep autoencoder Fault Detection Feature Extraction Random Forest Profile Extraction

来源：评论

学校读者我要写书评

暂无评论

deep autoencoder Neural Networks for Short-Term Traffic Congestion Prediction of Transportation Networks

引用

SENSORS 2019年第10期19卷 2229.-2229.页

作者： Zhang, Sen Yao, Yong Hu, Jie Zhao, Yong Li, Shaobo Hu, Jianjun Chinese Acad Sci Chengdu Inst Comp Applicat Chengdu 610041 Sichuan Peoples R China Univ Chinese Acad Sci Beijing 100049 Peoples R China Guizhou Inst Technol Sch Big Data Guiyang 550003 Guizhou Peoples R China Guizhou Univ Sch Mech Engn Guiyang 550003 Guizhou Peoples R China GuiZhou Univ Finance & Econ Coll Big Data Stat Guiyang 550025 Guizhou Peoples R China Univ South Carolina Dept Comp Sci & Engn Columbia SC 29201 USA

Traffic congestion prediction is critical for implementing intelligent transportation systems for improving the efficiency and capacity of transportation networks. However, despite its importance, traffic congestion prediction is severely less investigated compared to traffic flow prediction, which is partially due to the severe lack of large-scale high-quality traffic congestion data and advanced algorithms. This paper proposes an accessible and general workflow to acquire large-scale traffic congestion data and to create traffic congestion datasets based on image analysis. With this workflow we create a dataset named Seattle Area Traffic Congestion Status (SATCS) based on traffic congestion map snapshots from a publicly available online traffic service provider Washington State Department of Transportation. We then propose a deep autoencoder-based neural network model with symmetrical layers for the encoder and the decoder to learn temporal correlations of a transportation network and predicting traffic congestion. Our experimental results on the SATCS dataset show that the proposed DCPN model can efficiently and effectively learn temporal relationships of congestion levels of the transportation network for traffic congestion forecasting. Our method outperforms two other state-of-the-art neural network models in prediction performance, generalization capability, and computation efficiency.

关键词： transportation network traffic congestion forecasting spatial-temporal correlation deep learning end-to-end deep autoencoder convolutional neural network long short-term memory

来源：评论

学校读者我要写书评

暂无评论

A Novel Harris Hawk Algorithm (HHA) to Optimize deep autoencoder (DAE) Method of deep Learning for Cancer Diagnosis

A Novel Harris Hawk Algorithm (HHA) to Optimize Deep Autoenc...

引用

2023 Intelligent Computing and Control for Engineering and Business Systems, ICCEBS 2023

作者： Kumar, N. Jagadish Yuvaraj, G. Hameed, A. Thariq Noor, A. Mohamed Sheik Johindhan, E. Vasanth, P. Rudresh Sri Sai Ram Engineering College Department of Computer Science and Business Systems West Tambaram Chennai India West Tambaram Chennai India Velammal Institute of Technology Department of Information Technology Pancheti Chennai India

ISBN: (纸本)9798350394580

Over the past decade, the significance of data analytics in health informatics has surged due to the overwhelming influx of multi-modality data. Artificial neural networks have emerged as one of the most important and powerful influences in machine learning, rooted in deep learning. The exact prognostic of cancer remains a challenging problem for healthcare practitioners, even though cancer research has been a steady emphasis throughout the last decades. To increase a patient's chances of survival, early diagnosis is crucial, effectively differentiating between benign and malignant tumors is a considerable challenge. deep learning has become widely accepted tool among medical researchers as an effective method in this area. A deep autoencoder architecture which was primarily designed for feature extraction is improved by using the Harris Hawk algorithm. Surprisingly, this approach comprises a pre-training step rather than requiring labelled data for training. During this pre-training stage, the proposed Harris Hawk algorithm makes use of exploration, exploitation, and attacking techniques, increasing feature extraction accuracy significantly to an outstanding 93%. This in turn greatly improves the performance of deep auto encoder applications, including those for cell grouping, 3D- brain reconstruction, cancer detection, and the prediction of demographic data. The deep autoencoder optimized using the Harris Hawk method stood out as the obvious choice in terms of accuracy when compared to other autoencoder models including sparse autoencoders (SAE), denoising autoencoders (DAE), contractive autoencoders (CAE)and convolutional autoencoders (CNAE). © 2023 IEEE.

关键词： Cancer diagnosis deep autoencoder deep learning Feature extraction Optimization

来源：评论

学校读者我要写书评

暂无评论

Tasmanian devil optimization with deep autoencoder for intrusion detection in IoT assisted unmanned aerial vehicle networks

引用

AIN SHAMS ENGINEERING JOURNAL 2024年第11期15卷

作者： Negm, Noha Alamro, Hayam Allafi, Randa Khalid, Majdi Nouri, Amal M. Marzouk, Radwa Othman, Aladdin Yahya Ahmed, Noura Abdelaziz King Khalid Univ Coll Sci & Art Mahayil Dept Comp Sci Abha Saudi Arabia Princess Nourah Bint Abdulrahman Univ Coll Comp & Informat Sci Dept Informat Syst POB 84428 Riyadh 11671 Saudi Arabia Northern Border Univ Dept Comp & Informat Technol Coll Sci & Arts Ar Ar Saudi Arabia Umm Al Qura Univ Coll Comp & Informat Syst Dept Comp Sci Mecca Saudi Arabia Imam Abdulrahman Bin Faisal Univ Appl Coll Dept Comp Sci Dammam 34212 Saudi Arabia Prince Sattam Bin Abdulaziz Univ Dept Comp & Self Dev Preparatory Year Deanship Alkharj Saudi Arabia

Background: Recently, a developing count of physical objects is linked to the Internet at an unprecedented rate, calcifying the knowledge of the Internet of Things (IoT). In several paradigms of IoT applications, unmanned aerial vehicles (UAVs) and satellites for IoT have concerned much attention and are experiencing quick progress. As for UAVs, because of their superiority in maneuverability and cost, it is established an increasingly extensive consumption in several IoT scenarios like disaster relief, rapid transportation, and environment monitoring. Security remains a main problem in the IoT supported UAV networks that are solved by the employ of intrusion detection system (IDS) methods. Objective: This article aims to present a Tasmanian Devil Optimization with deep autoencoder for Intrusion Detection System (TDODAE-IDS) technique in IoT assisted Unmanned Aerial Vehicle Networks. Methods: The presented TDODAE-IDS technique majorly concentrates on the effectual identification of the intrusions in the IoT based UAV networks. To accomplish this, the presented TDODAE-IDS system designs a new TDO algorithm for the feature subset selection process. Moreover, the DAE model classifies the existence of intrusion in the UAV network and the hyperparameter tuning of the DAE model takes place using the dragonfly algorithm (DFA). Results: The simulation results of the TDODAE-IDS approach were tested on a benchmark IDS dataset and the results are assessed under several measures. Conclusion: The comprehensive comparative analysis highlighted the enhanced outcomes of the TDODAE-IDS algorithm over other recent approaches with maximum accuracy of 99.36%. Therefore, the proposed model can be employed to accomplish security in the IoT assisted UAV networks.

关键词： Internet of things Unmanned aerial vehicles deep autoencoder Intrusion detection Feature selection

来源：评论

学校读者我要写书评

暂无评论

A novel deep autoencoder and hyperparametric adaptive learning for imbalance intelligent fault diagnosis of rotating machinery

引用

ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE 2021年 102卷 104279-104279页

作者： Li, Wanxiang Shang, Zhiwu Gao, Maosheng Qian, Shiqi Zhang, Baoren Zhang, Jie Tianjin Modern Electromech Equipment Technol Key 399 Binshui West Rd Tianjin Peoples R China Tiangong Univ Sch Mech Engn 399 Binshui West Rd Tianjin Peoples R China

In engineering applications, normal signals are often much easier to obtain than fault signals, which makes training dataset appear class imbalance to affect the training of deep autoencoder. Besides, the anti-interference and robustness of the deep autoencoder are extremely susceptible to the influence of the noise environment, resulting in insufficient state feature extraction. Moreover, the selection of hyperparameter has no theoretical guidance, and needs a lot of experiments and manual experience to determine the appropriate parameters. To address the aforementioned shortcomings, this paper proposes a novel deep autoencoder and hyperparametric adaptive learning for intelligent fault diagnosis of rotating machinery. Firstly, a data augmentation strategy based on overlapping segmentation, rotation and jitter, called ORJ data augmentation strategy, is proposed to increase the data capacity of training samples and reduce the influence of class imbalance on deep autoencoder training. Then, a lossless constraint term and a weight nonnegative constraint term modification loss function are proposed to construct a deep sparse autoencoder with lossless and nonnegative constraints (SAE-LNC), which improves the robustness of the model and makes the extracted state features have higher quality. Finally, the artificial bee colony algorithm (ABC) is used to adaptively learn the hyperparameters of deep SAE-LNC to improve the intelligence and accuracy of the diagnostic model. The experimental results of vibration signals collected from bearing test-bed show that, compared with other diagnostic methods, the proposed method can effectively capture high-quality state features, enhance the robustness of feature extraction process and obtain higher diagnostic accuracy.

关键词： Fault diagnosis Rotating machinery deep autoencoder Artificial bee colony algorithm Adaptively learn Class imbalance

来源：评论

学校读者我要写书评

暂无评论

A Combination of deep autoencoder and Multi-Scale Residual Network for Landslide Susceptibility Evaluation

引用

REMOTE SENSING 2023年第3期15卷 653-653页

作者： Wang, Zhuolu Xu, Shenghua Liu, Jiping Wang, Yong Ma, Xinrui Jiang, Tao He, Xuan Han, Zeya Liaoning Tech Univ Sch Mapping & Geog Sci Fuxin 123000 Peoples R China Chinese Acad Surveying & Mapping Beijing 100830 Peoples R China Wuhan Univ Sch Resources & Environm Sci Wuhan 430072 Peoples R China

Landslide susceptibility evaluation can accurately predict the spatial distribution of potential landslides, which offers great usefulness for disaster prevention, disaster reduction, and land resource management. Aiming at the problems of insufficient samples for landslide compilation, difficulty in expanding landslide samples, and insufficient expression of nonlinear relationships among evaluation factors, this paper proposes a new evaluation method of landslide susceptibility combining deep autoencoder and multi-scale residual network (DAE-MRCNN). In the first step, a deep autoencoder network was used to learn the feature expression of the original landslide data in order to acquire effective features in the data. Next, a multi-scale residual network was constructed;specifically, the model was improved into a deep residual network model by adding skip connections in the convolutional layer. In addition, the multi-scale idea was utilized to fully extract the scale characteristics of the evaluation factors. Finally, the model was used for feature training, and the results were input into the Softmax classifier to complete the prediction of landslide susceptibility. For this purpose, a machine learning method and two state-of-the-art deep learning methods, namely SVM, CPCNN-ML, and 2D-CNN, were utilized to model landslide susceptibility in Hanzhong City, Shaanxi Province. The proposed method produced the highest model performance of 0.891, followed by 0.842, 0.869, and 0.873. The experimental results show that the DAE-MRCNN method can fully express the complex nonlinear relationships among the evaluation factors, alleviate the problem of insufficient samples in convolutional neural networks (CNN) training, and significantly improve the accuracy of susceptibility prediction.

关键词： landslide susceptibility deep autoencoder convolutional neural networks residual learning multiscale

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：