检索结果-内蒙古大学图书馆

Correntropy-Based Data Selective Adaptive Filtering

IEEE Transactions on Circuits and Systems I: Regular Papers 2024年第2期71卷 754-766页

作者： Chien, Ying-Ren Wu, Sheng-Teng Tsao, Hen-Wai Diniz, Paulo S.R. Adaptive and Autonomous Communication Laboratory Department of Electrical Engineering National Ilan University Yilan City260007 Taiwan College of Electrical Engineering and Computer Science Graduate Institute of Communication Engineering National Taiwan University Taipei10617 Taiwan Program of Electrical Engineering Department of Electronics and Computer Engineering COPPE/Poli Universidade Federal do Rio de Janeiro Rio de Janeiro68504 Brazil

Data selection can be used in conjunction with adaptive filtering algorithms to avoid unnecessary weight updating and thereby reduce computational overhead. This paper presents a novel correntropy-based data selection method as an alternative to conventional data selection mechanisms based on squared error values. We developed a variable correntropy sensing algorithm to maximize the instantaneous correntropy function for the Gaussian kernel function to mitigate the impact of impulse noise and other forms of noise that can be disregarded in data selection. The proposed data selection mechanism can be implemented with any adaptive filtering algorithm. In simulations, the proposed method (implemented with the least mean squared algorithm) outperformed comparable error-based data selection schemes in terms of hit rate and miss rate, and the resulting weight updating ratio was close to the expected weight updating ratio. © 2023 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.

关键词： Adaptive filters

来源：评论

学校读者我要写书评

暂无评论

A Speech Enhancement Algorithm Combining Wavelet Transform and Adaptive Filters 6

A Speech Enhancement Algorithm Combining Wavelet Transform a...

引用

6th IEEE International Conference on Image Processing, Applications and Systems, IPAS 2025

作者： Shrestha, Aayam Ghorshi, Seyed Joorabchi, Marjan Panahi, Issa Firouzeh, Fereshteh Fakhar The University of Texas at Tyler Department of Electrical and Computer Engineering TX United States Sharif University of Technology School of Science and Engineering Iran The University of Texas at Dallas Department of Electrical and Computer Engineering TX United States

ISBN: (纸本)9798331506520

Noise as an unwanted interference can significantly degrade speech signals, especially those recorded by many microphones. This interference is modeled as additive noise that originates from a range of sources including White Gaussian Noise (WGN), babble, crowd, large city, and traffic noises. These disturbances can alter the characteristics of speech signals reducing both their quality and intelligibility. This paper introduces a novel approach designed to reduce noise and enhance the quality and intelligibility of speech signals. The proposed method combines Wavelet Transform with Adaptive Filters, specifically the Wiener filter and RLS filter. The evaluation process involves testing noisy speech signals under realistic conditions with different signal-to-noise ratios (SNRs) and different types of additive noise. The objective measure is used for evaluation, including the perceptual evaluation of speech quality (PESQ). Results show that combining Wiener or RLS filtering with Wavelet Transform significantly improves noise reduction, outperforming the use of Wavelet Transform alone. © 2025 IEEE.

关键词： Adaptive filtering

来源：评论

学校读者我要写书评

暂无评论

Alzheimer's Disease Stage Classification using Blood Data 2

Alzheimer's Disease Stage Classification using Blood Data

引用

2nd International Conference on Artificial Intelligence, Blockchain, and Internet of Things, AIBThings 2024

作者： Rasheed, Asif Fadlullah, Zubair Md Fouda, Mostafa M. Lakehead University Department of Computer Science Thunder Bay Canada Western University Department of Computer Science London Canada Idaho State University Department of Electrical and Computer Engineering Pocatello United States

ISBN: (纸本)9798331529529

In this work, we aim to evaluate the performance of Machine Learning models in the classification of Alzheimer's patients into disease stages using two feature selection methods proposed in our previous work. The first method identifies relevant features by ranking them according to their dependency on the diagnosis, which is measured using metrics such as Mutual Information, Symmetric Uncertainty and Cramer's V. The second method filters relevant features by applying a threshold to the Euclidean distance between the class means for each feature. Three new approaches are proposed to accommodate multiple classes in the Euclidean distance-based feature selection method. These approaches involve scoring each feature using the minimum, maximum and average of the Euclidean distances between the class means. Using two datasets, we extensively tested all possible combinations of relevant features identified using the two methods, ensuring minimal correlation among features in each panel. The panels achieved accuracies as high as 75.17% and 85.46% on the first and second datasets, respectively. The performances achieved in this work demonstrate the feasibility of using blood-based biomarkers in Alzheimer's disease stage classification and could help in developing accurate low-cost and non-invasive early detection tools that could aid in Alzheimer's disease diagnosis. © 2024 IEEE.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Effective NSGA-II Algorithm for a Limited AGV Scheduling Problem in Matrix Manufacturing Workshops with Undirected Material Flow

引用

Complex System Modeling and Simulation 2025年第1期5卷 68-85页

作者： Xuewu Wang Jianing Zhang Yi Hua Rui Yu School of Information Science and Engineering East China University of Science and TechnologyShanghai 200237China Institute for Sustainable Manufacturing and also with the Department of Electrical and Computer EngineeringUniversity of KentuckyLexingtonKY 40506USA

Automatic guided vehicles(AGVs)are extensively employed in manufacturing workshops for their high degree of automation and *** paper investigates a limited AGV scheduling problem(LAGVSP)in matrix manufacturing workshops with undirected material flow,aiming to minimize both total task delay time and total task completion *** address this LAGVSP,a mixed-integer linear programming model is built,and a nondominated sorting genetic algorithm II based on dual population co-evolution(NSGA-IIDPC)is *** NSGA-IIDPC,a single population is divided into a common population and an elite population,and they adopt different evolutionary strategies during the evolution *** dual population co-evolution mechanism is designed to accelerate the convergence of the non-dominated solution set in the population to the Pareto front through information exchange and competition between the two *** addition,to enhance the quality of initial population,a minimum cost function strategy based on load balancing is *** local search operators based on ideal point are proposed to find a better local *** improve the global exploration ability of the algorithm,a dual population restart mechanism is *** tests and comparisons with other algorithms are conducted to demonstrate the effectiveness of NSGA-IIDPC in solving the LAGVSP.

关键词： limited automatic guided vehicle(AGV)scheduling problem nondominated sorting genetic algorithm II(NSGA-II) dual population co-evolution matrix manufacturing workshop

来源：评论

学校读者我要写书评

暂无评论

Keep in Balance: Runtime-reconfigurable Intermittent Deep Inference

引用

ACM Transactions on Embedded Computing Systems 2023年第5 s期22卷 1-25页

作者： Yen, Chih-Hsuan Mendis, Hashan Roshantha Kuo, Tei-Wei Hsiu, Pi-Cheng Department of Computer Science and Information Engineering National Taiwan University Taiwan Academia Sinica Taiwan Mohamed Bin Zayed University of Artificial Intelligence United Arab Emirates College of Electrical Engineering and Computer Science National Taiwan University Taiwan Department of Computer Science and Information Engineering National Chi Nan University Taiwan

Intermittent deep neural network (DNN) inference is a promising technique to enable intelligent applications on tiny devices powered by ambient energy sources. Nonetheless, intermittent execution presents inherent challenges, primarily involving accumulating progress across power cycles and having to refetch volatile data lost due to power loss in each power cycle. Existing approaches typically optimize the inference configuration to maximize data reuse. However, we observe that such a fixed configuration may be significantly inefficient due to the fluctuating balance point between data reuse and data refetch caused by the dynamic nature of ambient *** work proposes DynBal, an approach to dynamically reconfigure the inference engine at runtime. DynBal is realized as a middleware plugin that improves inference performance by exploring the interplay between data reuse and data refetch to maintain their balance with respect to the changing level of intermittency. An indirect metric is developed to easily evaluate an inference configuration considering the variability in intermittency, and a lightweight reconfiguration algorithm is employed to efficiently optimize the configuration at runtime. We evaluate the improvement brought by integrating DynBal into a recent intermittent inference approach that uses a fixed configuration. Evaluations were conducted on a Texas Instruments device with various network models and under varied intermittent power strengths. Our experimental results demonstrate that DynBal can speed up intermittent inference by 3.26 times, achieving a greater improvement for a large network under high intermittency and a large gap between memory and computation performance. © 2023 Copyright held by the owner/author(s). Publication rights licensed to ACM.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

Detecting COVID-19 and Lung Diseases Utilizing Parallel CNN Model on Chest X-Rays Enhanced with GradCAM Visualization 2

Detecting COVID-19 and Lung Diseases Utilizing Parallel CNN ...

引用

2nd International Conference on Power, electrical, Electronics and Industrial Applications, PEEIACON 2024

作者： Kibria, Hafsa Binte Hossain, Md. Ali Rajshahi University of Engineering & Technology Department of Electrical & Computer Engineering Rajshahi6204 Bangladesh Rajshahi University of Engineering & Technology Department of Computer Science Engineering Rajshahi6204 Bangladesh

ISBN: (纸本)9798331517984

The emergence of COVID-19 has underscored the urgency of accurate medical diagnosis, particularly in the context of chest X-ray image classification for various lung conditions, including COVID-19, normal cases, viral pneumonia, and lung opacity. In light of the pandemic's impact, there is a pressing need to develop resource-efficient solutions for automated classification, specifically tailored for resource-constrained devices. This article addresses this need by proposing a Lightweight Parallel CNN (Convolutional Neural Network) architecture capable of effectively categorizing the aforementioned multi-class cases. Leveraging a comprehensive dataset of 21165 images, our model achieves a decent accuracy of 93.21% for four-class classification. Notably, the proposed architecture maintains efficiency with fewer than 783k parameters, demonstrating its potential for deployment on devices with limited computational resources. This research contributes to the advancement of accessible and accurate medical diagnostics, catering to the demand for streamlined and reliable solutions amidst challenging circumstances. © 2024 IEEE.

关键词： COVID-19

来源：评论

学校读者我要写书评

暂无评论

Coprocessor Actor Critic: A Model-Based Reinforcement Learning Approach For Adaptive Brain Stimulation 41

Coprocessor Actor Critic: A Model-Based Reinforcement Learni...

引用

41st International Conference on Machine Learning, ICML 2024

作者： Pan, Michelle Schrum, Mariah Myers, Vivek Bıyık, Erdem Dragan, Anca Department of Electrical Engineering and Computer Sciences UC Berkeley United States Department of Computer Science University of Southern California United States

Adaptive brain stimulation can treat neurological conditions such as Parkinson's disease and post-stroke motor deficits by influencing abnormal neural activity. Because of patient heterogeneity, each patient requires a unique stimulation policy to achieve optimal neural responses. Model-free reinforcement learning (MFRL) holds promise in learning effective policies for a variety of similar control tasks, but is limited in domains like brain stimulation by a need for numerous costly environment interactions. In this work we introduce Coprocessor Actor Critic, a novel, model-based reinforcement learning (MBRL) approach for learning neural coprocessor policies for brain stimulation. Our key insight is that coprocessor policy learning is a combination of learning how to act optimally in the world and learning how to induce optimal actions in the world through stimulation of an injured brain. We show that our approach overcomes the limitations of traditional MFRL methods in terms of sample efficiency and task success and outperforms baseline MBRL approaches in a neurologically realistic model of an injured brain. Copyright 2024 by the author(s)

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Investigation Emotion Dynamics in Bangla Social Media Text for Improved Emotion Detection

Investigation Emotion Dynamics in Bangla Social Media Text f...

引用

2024 IEEE International Conference on Computing, Applications and Systems, COMPAS 2024

作者： Ahmed, Mostak Sobhan, Abdus Hossen Sajib, Md. Sahadat Masud, Nasrullah Green University of Bangladesh Department of Computer Science and Engineering Dhaka Bangladesh Varendra University Department of Computer Science and Engineering Rajshahi Bangladesh Varendra University Department of Electrical and Electronic Engineering Rajshahi Bangladesh

ISBN: (纸本)9798331529765

Recognizing emotion from text is fundamental to machine learning and influences our understanding of human interaction. While English sentiment analysis has been well-researched, Bengali (Bangla) is still under-researched. Our study bridges this gap by investigating the emotional dynamics in Bengali social media posts using extensive user-generated content. We use a modified dataset defined by seven emotions: Happy, Surprise, Fearful, Angry, Neutral, Disgust, and Sad. Using machine learning techniques such as support vector classifiers, decision trees, logistic regression, and random forest, we found that the Random Forest model achieved the highest accuracy of 84%. This study advances our understanding of Bengali emotional expression and demonstrates its application in social media sentiment analysis, psychoanalysis, and communication technologies. By addressing the complexities of Bengali language and culture, our work contributes to a broader definition of emotion and lays the foundation for future research in different contexts. © 2024 IEEE.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Broadband spin and angle co‑multiplexed waveguide‑based metasurface for six‑channel crosstalk‑free holographic projection

引用

eLight 2024年第1期4卷 223-237页

作者： Zeyang Liu Hao Gao Taigao Ma Vishva Ray Niu Liu Xinliang Zhang L.Jay Guo Cheng Zhang School of Optical and Electronic Information Huazhong University of Science and TechnologyWuhan 430074China Wuhan National Laboratory for Optoelectronics Huazhong University of Science and TechnologyWuhan 430074China Department of Physics University of MichiganAnn ArborMI 48109USA Lurie Nanofabrication Facility Department of Electrical Engineering and Computer ScienceUniversity of MichiganAnn ArborMI 48109USA Department of Electrical Engineering and Computer Science University of MichiganAnn ArborMI 48109USA

Metasurface-based holograms,or metaholograms,offer unique advantages including enhanced imaging quality,expanded field of view,compact system size,and broad operational ***-channel metaholograms,capable of switching between multiple projected images based on the properties of illuminating light such as state of polarization and angle of incidence,have emerged as a promising solution for realizing switchable and dynamic holographic ***,existing designs typically grapple with challenges such as limited multiplexing channels and unwanted crosstalk,which severely constrain their practical ***,we present a new type of waveguidebased multi-channel metaholograms,which support six independent and fully crosstalk-free holographic display channels,simultaneously multiplexed by the spin and angle of guided incident light within the glass *** employ a k-space translation strategy that allows each of the six distinct target images to be selectively translated from evanescent-wave region to the center of propagation-wave region and projected into free space without crosstalk,when the metahologram is under illumination of a guided light with specific spin and azimuthal *** addition,by tailoring the encoded target images,we implement a three-channel polarization-independent metahologram and a two-channel full-color(RGB)***,the number of multiplexing channels can be further increased by expanding the k-space’s central-period region or combing the k-space translation strategy with other multiplexing techniques such as orbital angular momentum *** work provides a novel approach towards realization of high-performance and compact holographic optical elements with substantial information capacity,opening avenues for applications in AR/VR displays,image encryption,and information storage.

关键词： Metasurface Multi-channel metahologram Color metahologram Waveguide display

来源：评论

学校读者我要写书评

暂无评论

Robust Defense Against Extreme Grid Events Using Dual-Policy Reinforcement Learning Agents

Robust Defense Against Extreme Grid Events Using Dual-Policy...

引用

2025 IEEE Texas Power and Energy Conference, TPEC 2025

作者： Peter, Benjamin M. Korkali, Mert University of Missouri Department of Electrical Engineering and Computer Science ColumbiaMO65211 United States

ISBN: (纸本)9798331541125

Reinforcement learning (RL) agents are powerful tools for managing power grids. They use large amounts of data to inform their actions and receive rewards or penalties as feedback to learn favorable responses for the system. Once trained, these agents can efficiently make decisions that would be too computationally complex for a human operator. This ability is especially valuable in decarbonizing power networks, where the demand for RL agents is increasing. These agents are well suited to control grid actions since the action space is constantly growing due to uncertainties in renewable generation, microgrid integration, and cybersecurity threats. To assess the efficacy of RL agents in response to an adverse grid event, we use the Grid2Op platform for agent training. We employ a proximal policy optimization (PPO) algorithm in conjunction with graph neural networks (GNNs). By simulating agents' responses to grid events, we assess their performance in avoiding grid failure for as long as possible. The performance of an agent is expressed concisely through its reward function, which helps the agent learn the most optimal ways to reconfigure a grid's topology amidst certain events. To model multi-actor scenarios that threaten modern power networks, particularly those resulting from cyberattacks, we integrate an opponent that acts iteratively against a given agent. This interplay between the RL agent and opponent is utilized in N-k contingency screening, providing a novel alternative to the traditional security assessment. © 2025 IEEE.

关键词： Graph neural networks

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：