检索结果-内蒙古大学图书馆

25th International Conference on Speech and Computer (SPECOM)

作者： Patel, Anuj Prasad, G. Satya Chandra, Sabyasachi Bharati, Puja Das Mandal, Shyamal Kumar Indian Inst Technol Kharagpur W Bengal India Pandit Deendayal Energy Univ Gandhinagar India

ISBN: (纸本)9783031483080;9783031483097

Speech enhancement techniques play a vital role in enhancing the clarity and overall quality of audio signals, addressing issues like background noise, reverberation, and channel impairments that often degrade speech intelligibility. Neural network models, including DNNs, CNNs, RNNs, and VAEs, have demonstrated their effectiveness in improving speech quality by decoding noisy speech inputs, capturing intricate patterns, and extracting relevant information. Evaluation metrics like PESQ and STOI are commonly employed to assess the performance of speech enhancement algorithms. STOI measures the understandability of enhanced speech using short-time spectral information, while PESQ evaluates the subjective quality of enhanced speech compared to the original clean speech. Moreover, recent advancements in speech enhancement research have shown that employing LinkNet, a specific neural network architecture, can significantly surpass the efficiency of other models. LinkNet has demonstrated superior performance in enhancing speech signals by effectively mitigating noise, reducing artifacts, and enhancing the overall intelligibility of the output. Its architecture incorporates innovative techniques that facilitate the extraction of meaningful features from noisy speech inputs, leading to remarkable results in terms of speech quality improvement. By leveraging LinkNet, researchers and practitioners can further advance the field of speech enhancement and achieve outstanding outcomes in terms of audio clarity and intelligibility.

关键词： Speech enhancement Deep learning Convolutional neural networks auto encoder UNET LinkNet-speech

来源：评论

学校读者我要写书评

暂无评论

Intrusion Detection based on extracting Optimization Features for Bidirectional Long-Short-Term-Memory 24

Intrusion Detection based on extracting Optimization Feature...

引用

9th International Conference on Intelligent Information Technology (ICIIT)

作者： Hoang Thien Van Phuoc Hong Minh Trong Nguyen HUTECH Univ Fac Informat Technol Ho Chi Minh City Vietnam Vietnam Natl Univ Univ Informat Technol UIT Ho Chi Minh City Vietnam

ISBN: (纸本)9798400716713

In today's insecure digital landscape, the need for robust Intrusion Detection Systems (IDS) is more critical than ever. Network attacks and unauthorized intrusions pose significant threats, requiring effective IDS development. Many intrusion detection algorithms combine traditional machine learning models with deep learning to enhance performance, but they still encounter issues like insufficient detection accuracy and potential accuracy reduction due to data preprocessing operations. This paper introduces a novel intrusion detection method that enhances Bidirectional Long-Short-Term-Memory (BiLSTM) performance using Logarithmic autoencoder (LogAE) and Extreme Gradient Boosting (XGBoost) for optimal feature extraction, aiming for high detection rates. LogAE generates new features from input feature sets, which are then combined with the original features. XGBoost is employed to identify key features from this combined set, which serves as input to BiLSTM for robust attack classification. The proposed algorithms are evaluated on the NSL-KDD dataset using various evaluation metrics, demonstrating the effectiveness of the proposed method in improving IDS performance and providing valuable insights for secure and efficient intrusion detection systems against cyber security threats.

关键词： Intrusion Detection System Network Security Machine Learning auto encoder Extreme Gradient Boosting Bidirectional Long-Short-Term-Memory

来源：评论

学校读者我要写书评

暂无评论

Parkinson’s Detection Using RNN-Graph-LSTM with Optimization Based on Speech Signals

引用

Computers, Materials & Continua 2022年第7期72卷 871-886页

作者： Ahmed S.Almasoud Taiseer Abdalla Elfadil Eisa Fahd N.Al-Wesabi Abubakar Elsafi Mesfer Al Duhayyim Ishfaq Yaseen Manar Ahmed Hamza Abdelwahed Motwakel Department of Information Systems College of Computer and Information SciencesPrince Sultan UniversityMahayil62529Saudi Arabia Department of Information Systems-Girls Section King Khalid UniversityMahayil62529Saudi Arabia Department of Computer Science College of Science and Artsat MahayilKing Khalid UniversityMahayil62529Saudi Arabia Faculty of Computer and IT Sana’a UniversitySana’aYemen Department of Software Engineering College of Computer Science and EngineeringUniversity of JeddahSaudi Arabia Department of Natural and Applied Sciences College of Community-AflajPrince Sattam bin Abdulaziz UniversityMahayil62529Saudi Arabia Department of Computer and Self Development Preparatory Year DeanshipPrince Sattam bin Abdulaziz UniversityAlKharj62529Saudi Arabia

Early detection of Parkinson’s Disease(PD)using the PD patients’voice changes would avoid the intervention before the identification of physical *** machine learning algorithms were developed to detect PD ***,these ML methods are lack in generalization and reduced classification performance due to subject *** overcome these issues,this proposed work apply graph long short term memory(GLSTM)model to classify the dynamic features of the PD patient speech *** proposed classification model has been further improved by implementing the recurrent neural network(RNN)in batch normalization layer of GLSTM and optimized with adaptive moment estimation(ADAM)on network hidden *** consider the importance of feature engineering,this proposed system use Linear Discriminant analysis(LDA)for dimensionality reduction and Sparseauto-encoder(SAE)for extracting the dynamic speech *** on the computation of energy content transited from unvoiced to voice(onset)and voice to voiceless(offset),dynamic features are *** PD datasets is evaluated under 10 fold cross validation without sample *** proposed smart PD detection method called RNN-GLSTM-ADAM is numerically experimented with persistent phonations in terms of accuracy,sensitivity,and specificity andMatthew correlation *** evaluated result of RNN-GLSTM-ADAM extremely improves the PD detection accuracy than static feature based conventional ML and DL approaches.

关键词： Dimensionality reduction LDA feature extraction auto encoder classification recurrent neural network LSTM and optimization

来源：评论

学校读者我要写书评

暂无评论

False Positive Elimination in Object Detection Methods for Videos 16

False Positive Elimination in Object Detection Methods for V...

引用

16th International Conference on Machine Vision (ICMV)

作者： Dubey, Shubham Kumar Satyanarayana, J. V. Mohan, C. Krishna Indian Inst Technol Hyderabad Dept Comp Sci Hyderabad India Indian Inst Technol Hyderabad Hyderabad India

ISBN: (纸本)9781510674639;9781510674622

A robust object detection algorithm is essential while detecting objects in videos and real time scenarios, where false positives might result in unwanted outcomes. Our goal here is to observe how Simple Online and Real-time Tracking with a Deep association metric (Deep SORT) algorithm for Multi-Object Tracking (MOT) can be used to minimize false positives, from a state of the art detection algorithm like You Only Look Once (YOLO), by using the Kalman filter approach. An auto encoder based feature extractor has been used, instead of the standard CNN networks like ResNet-50 to further improve speed of the detector. There have been other MOT algorithms in the recent times which give good results, but are not as real time efficient as the simple yet efficient Deep SORT method. Experimental analysis has shown how autoencoder based Deep SORT performs in contrast to native Deep SORT and YOLO, in eliminating false positive detections.

关键词： Object Detection Deep SORT False positive Kalman Filter auto encoder

来源：评论

学校读者我要写书评

暂无评论

A state-classification approach for light-weight robotic drilling using model-based data augmentation and multi-level deep learning

引用

MECHANICAL SYSTEMS AND SIGNAL PROCESSING 2022年第PartA期167卷 108480-108480页

作者： Lu, Hao Zhao, Xingwei Tao, Bo Ding, Han Huazhong Univ Sci & Technol Sch Mech Sci & Engn State Key Lab Digital Mfg Equipment & Technol Wuhan 430074 Hubei Peoples R China

In the aircraft assembly industry, a large number of holes need to be drilled so that a convenient and fast drilling quality detection method is necessary. During the robotic drilling process, the verticality of the drill directly affects the quality of the hole. In this paper, a vibration-based classification approach is proposed to find out the unqualified holes caused by inclined drilling. Firstly, robotic drilling's vibration model is established to simulate vibration signals during vertical and inclined drilling, which to prove the model information of the drill system. Then, based on auto encoder, the actual drilling signals are classified into drilling section and non drilling section, and the stable drilling section signals are intercepted from the drilling section to be further used to improve classification performance. Finally, combining the simulated signals and a small number of actual signals as the training set, the data augmentation is realized and a deep residual network (Resnet) classifier is trained. Inputting the stable drilling section signals into the classifier, the drilling state is classified into vertical drilling and inclined drilling. In order to verify the proposed method, sixteen groups of experiments under different process parameters are carried out in this paper, and the results have proved the effectiveness and potential of this method.

关键词： Machining signal processing Robotic drilling Data augmentation Vibration model auto encoder Deep residual network

来源：评论

学校读者我要写书评

暂无评论

Deep Learning based Network based Intrusion Detection System in Industrial Internet of Things 2

Deep Learning based Network based Intrusion Detection System...

引用

2nd International Conference on Intelligent Data Communication Technologies and Internet of Things, IDCIoT 2024

作者： Beauty Angelin, J. Alwina Priyadharsini, C. Karunya Institute of Technology and Sciences Department of Computer Science and Engineering Coimbatore India

ISBN: (纸本)9798350327533

The Industrial Internet of Things (IIoT) is the result of integrating the Internet of Things (IoT) into critical industries, especially in industrial and production settings. The industrial Internet of things (IoT) has many benefits, but security and privacy remain major concerns. The current intrusion detection systems (IDS) for the Internet of things have encountered a number of problems, including the inability to handle different kinds of attacks, the dependence on traditional datasets, and the lack of attention to imbalanced datasets. In addressing these challenges, this study proposes the implementation of a deep learning-based network intrusion detection system designed to identify diverse types of attacks within industrial Internet of Things environments. And to improve the effectiveness of the Network Intrusion Detection System (NIDS), the suggested model combines Convolutional Neural Network (CNN) with auto encoder (AE) techniques. The model improves detection outcomes by reducing data features using an autoencoder. Additionally, Convolutional Neural Networks possess the ability to automatically extract intricate features and patterns from complex data, such as network traffic. The hybrid model exhibits commendable performance on the Edge_IIoT dataset, achieving an accuracy of 92.34%, precision of 91.69%, recall of 90.28%, and an F1 score of 89.08%. © 2024 IEEE.

关键词： auto encoder Convolutional Neural Network Cyberattacks Deep learning Intrusion detection system

来源：评论

学校读者我要写书评

暂无评论

Dimension Reduction Techniques for Optimization 1

Dimension Reduction Techniques for Optimization

引用

1st International Conference on Computing, Sciences and Communications, ICCSC 2024

作者： Saxena, Abhishek Kumar Pandey, Rajiv Singh, Neeraj Kumar AIIT Amity University Lucknow India University of Toulouse INPT-ENSEEIHT/IRIT France

ISBN: (纸本)9798350353648

Dimension-reduction techniques can be used to optimize a class problems. We investigate various dimensional reduction techniques to find the best method for our data processing. It is thoroughly discussed how features are extracted and selected. We discuss the various factors that must be considered when reducing dimensions in this paper. In the later part of this paper, we discuss in detail the basic techniques such as principal component analysis, kernel principal component analysis and auto encoder. © 2024 IEEE.

关键词： and feature extraction auto encoder Dimension reduction factor analysis feature selection Google colab PCA

来源：评论

学校读者我要写书评

暂无评论

MOLECULES GENERATION TO INHIBIT COVID-19 DISEASE USING encoder-DECODER LSTM ARCHITECTURE AND PCA PROPERTIES

引用

JOURNAL OF MECHANICS IN MEDICINE AND BIOLOGY 2023年第1期23卷

作者： Amrane, Meriem Oukid, Salyha Ensari, Tolga Univ Saad Dahlab Blida Comp Sci Dept Blida Algeria Arkansas Tech Univ Dept Comp & Informat Sci Russellville AR 72801 USA

COVID-19 has become the world's worst pandemic and has claimed over six million lives as of March 2022. The virus is now in alongside cancer as one of the most common causes of death. Likewise, there is no definitive or unique treatment for COVID-19 outside of a selected few drugs approved by the Food and Drug Administration (FDA). While Artificial Intelligence (AI) can be used to generate molecules that target Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2), the virus responsible for COVID-19, such molecules are novel and do not yet exist in the market. With the emergence and availability of several drug datasets related to COVID-19 (tests, images, graphs, and ChEMBLs), recent works based on Deep Learning (DL) techniques have been employed to generate molecules and check the effectiveness of existing molecules on COVID-19. In our study, we investigated the benefits of an encoder-Decoder (ED) architecture based on Long Short-Term Memory (LSTM) cells. As a result, the molecules were converted into a vector during the encoding phase, which was then decoded back into SMILES molecules strings. We propose an approach to incorporate four features of Principal Components Analysis (PCA) with encoder-Decoder Long Short-Term Memory (ED-LSTM) for regularization, which means that, instead of avoiding linear mapping, we assumed that the data could be linearly separable. We concluded that ED-LSTM with unit norm constraint has the best reconstruction accuracy in the context of generating molecules. The resulting dataset was used with the aid of virtual screening and convolutional neural networks to check the drugs that have the best binding affinity with SARS-CoV-2. We achieved an accuracy of 87.35% on the test set.

关键词： auto encoder molecules generation deep learning LSTM drug discovery COVID-19 SMILES

来源：评论

学校读者我要写书评

暂无评论

Construction of business innovation model for sports industry using a deep learning algorithm

引用

SOFT COMPUTING 2022年第20期26卷 10753-10763页

作者： Lv, Chenchen Wang, Yifeng Ma, Yin Xidian Univ Sch Econ & Management Xian 710126 Shaanxi Peoples R China Xian Phys Educ Univ Sch Sports Econ & Management Xian 710068 Shaanxi Peoples R China

The rapid development and advancement in Internet technology have provided a common interface to different sporting companies to use the Internet as a platform to develop business models for the sports industry. The number of players and supporting staff increases with the passage of time and they generate data with a much higher speed, resulting in a large volume of sports-related data. The analysis and forecasting of such a huge quantities of data is a great task that cannot be solved using standard methods. To handle this issue, we designed and developed an intelligent deep learning (DL) model based on the concept of stacked self-coding network, which uses a DL framework for decision making at the backend. The proposed model can learn and select the embedding of original features in the high dimensional, sparse, and noisy big data environment of the sports industry. The dataset used for the analysis and implementation purpose consists of a total of 2,053,500 samples, which is further divided into six sample sets by introducing random noise of different intensities to the test sample set. The performance of the proposed model in risk assessment and feature embedding learning is also described from a visual analysis perspective. Up to a point, the interpretability of DL technology in the sports business is also investigated and examined. The experimental results of credit risk assessment attained via the proposed model are compared with the traditional method, which verifies the effectiveness and significance of the proposed model. In terms of noise decrease, the exploratory comes approximately appears that the proposed model encompasses a higher disturbance resistance capability towards huge information in a loud environment, as well as for the input information with noise obstructions.

关键词： Sports industry Internet Deep learning Decision-making auto encoder Risk assessment

来源：评论

学校读者我要写书评

暂无评论

MARL: Multi-scale Archetype Representation Learning for Urban Building Energy Modeling

MARL: Multi-scale Archetype Representation Learning for Urba...

引用

IEEE/CVF International Conference on Computer Vision (ICCV)

作者： Zhuang, Xinwei Huang, Zixun Zeng, Wentao Caldas, Luisa Univ Calif Berkeley Berkeley CA 94720 USA

ISBN: (纸本)9798350307443

Building archetypes, representative models of building stock, are crucial for precise energy simulations in Urban Building Energy Modeling. The current widely adopted building archetypes are developed on a nationwide scale, potentially neglecting the impact of local buildings' geometric specificities. We present Multi-scale Archetype Representation Learning (MARL), an approach that leverages representation learning to extract geometric features from a specific building stock. Built upon VQ-AE, MARL encodes building footprints and purifies geometric information into latent vectors constrained by multiple architectural downstream tasks. These tailored representations are proven valuable for further clustering and building energy modeling. The advantages of our algorithm are its adaptability with respect to the different building footprint sizes, the ability for automatic generation across multi-scale regions, and the preservation of geometric features across neighborhoods and local ecologies. In our study spanning five regions in LA County, we show MARL surpasses both conventional and VQ-AE extracted archetypes in performance. Results demonstrate that geometric feature embeddings significantly improve the accuracy and reliability of energy consumption estimates. Code, dataset and trained models are available on the project page: https://***/ZixunHuang1997/MARLBuildingEnergyEstimation.

关键词： auto encoder Building archetype Representation learning Urban building energy modeling

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：