检索结果-内蒙古大学图书馆

Enhanced battery life prediction with reduced data demand via semi-supervised representation learning

Journal of Energy Chemistry 2025年第2期101卷 524-534,I0011页

作者： Liang Ma Jinpeng Tian Tieling Zhang Qinghua Guo Chi Yung Chung School of Mechanical MaterialsMechatronic and Biomedical EngineeringUniversity of WollongongNSW 2522Australia Department of Electrical and Electronic Engineering and Research Centre for Grid Modernisation The Hong Kong Polytechnic UniversityKowloonHong Kong 999077China Photonics Research Institute The Hong Kong Polytechnic UniversityKowloonHong Kong 999077China School of Electrical Computer and Telecommunications EngineeringUniversity of WollongongWollongongNSW 2522Australia

Accurate prediction of the remaining useful life(RUL)is crucial for the design and management of lithium-ion *** various machine learning models offer promising predictions,one critical but often overlooked challenge is their demand for considerable run-to-failure data for *** of such training data leads to prohibitive testing efforts as the run-to-failure tests can last for ***,we propose a semi-supervised representation learning method to enhance prediction accuracy by learning from data without RUL *** approach builds on a sophisticated deep neural network that comprises an encoder and three decoder heads to extract time-dependent representation features from short-term battery operating data regardless of the existence of RUL *** approach is validated using three datasets collected from 34 batteries operating under various conditions,encompassing over 19,900 charge and discharge *** method achieves a root mean squared error(RMSE)within 25 cycles,even when only 1/50 of the training dataset is labelled,representing a reduction of 48%compared to the conventional *** also demonstrate the method's robustness with varying numbers of labelled data and different weights assigned to the three decoder *** projection of extracted features in low space reveals that our method effectively learns degradation features from unlabelled *** approach highlights the promise of utilising semi-supervised learning to reduce the data demand for reliability monitoring of energy devices.

关键词： Lithium-ion batteries Battery degradation Remaining useful life Semi-supervised learning

来源：评论

学校读者我要写书评

暂无评论

Creating a Framework for Non-Profit Impact Reporting

Creating a Framework for Non-Profit Impact Reporting

引用

Annual IEEE Systems Conference

作者： Parker Langa Mona Blake Sarah Fenton Michael Nixdorf Shamsnaz V. Bhada Electrical and Computer Engineering Worcester Polytechnic Institute Worcester MA Mechanical Engineering Worcester Polytechnic Institute Worcester MA Chemical Engineering Worcester Polytechnic Institute Worcester MA Theoretical Physics Worcester Polytechnic Institute Worcester MA

ISBN: (数字)9798331508180

ISBN: (纸本)9798331508197

Non-profit organizations are increasingly pressured to provide impact analysis highlighting their community influence to continue receiving financial support. Organizations with limited resources can struggle in areas such as data organization, visualization, and conflicting stakeholders' expertise and interest. The need to show evidence-based impact while struggling with resources is a balancing act of survival for many non-profits and a unique case of a Socio-Technical system. In our research, we attempt to tackle this Socio-Technical problem by creating a framework and building tools to alleviate this balancing act to allow non-profits to devote their resources to supporting their mission. We use a mixed methods approach through data collection and analysis to provide a cohesive overview of a charity's influence while requiring few resources. We apply this method to Historic Buildings & Places (HB&P), a non-profit organization in the UK that works to preserve built heritage to retain its cultural value and increase sustainable infrastructure. Using HB&P as a case study provides an understanding of practices used to complete this mixed methods approach. This paper describes the process for determining impact criteria and creating an accessible report as well as providing a set of terms and a procedure applicable to non-profits across many sectors. The findings include which graphics wider audiences easily understand—column graphs and maps—and which themes are currently sought after—diversity and sustainability. Given our interview pool focused on current and potential stakeholders, non-profits who focus on highlighting their diversity and sustainability work through accessible graphics have the potential to increase their financial support.

关键词： Sociotechnical systems Visualization Data analysis Buildings Software Stakeholders Sustainable development Interviews Heat maps Lenses

来源：评论

学校读者我要写书评

暂无评论

Efficient Path Selection Design for Large Scale LEO Satellite Constellations Using Graph Embedding-Based Reinforcement Learning

引用

IEEE Transactions on Network Science and engineering 2025年第3期12卷 2007-2020页

作者： Kang, Yuhan Zhu, Yifei Wang, Dan Han, Zhu University of Houston Electrical and Computer Engineering Department HoustonTX77004 United States Shanghai Jiao Tong University UM-SJTU Joint Institute Shanghai200240 China The Hong Kong Polytechnic University Department of Computing Hong Kong University of Houston Department of Electrical and Computer Engineering HoustonTX77004 United States Kyung Hee University Department of Computer Science and Engineering Seoul446-701 Korea Republic of

The rapid expansion of large-scale Low Earth Orbit (LEO) satellite constellations marks a new phase in global connectivity and communication. However, efficient path selection among numerous satellites connected by inter-satellite links (ISL) in such dynamic networks poses substantial challenges. This paper introduces a novel Path Selection Mechanism (PSM-LEO) utilizing Graph Embedding-Based Reinforcement Learning (GERL). Our GERL-based PSM-LEO method employs Graph Embedding (GE) to model the satellite network in a simplified low-dimensional space, simplifying the analysis of complex satellite relationships. Combining Reinforcement Learning (RL) with GE allows for dynamic adaptation of path choices in response to dynamic network conditions and communication needs. To the best of our knowledge, we are the first to propose GERL for PSM-LEO. We evaluate our method's performances through simulations in the Ansys Systems Tool Kit (STK), focusing on diverse LEO scenarios. Our results reveal that our approach surpasses several benchmarks in convergence speed, end-to-end latency, and energy consumption, demonstrating enhanced data transfer efficiency, reduced latency, and improved reliability. Additionally, we assess the scalability of the proposed method by analyzing its performance with increasing satellite constellation sizes. Our results confirm the framework's high scalability, demonstrating its suitability for addressing path selection challenges in future larger satellite networks. © 2013 IEEE.

关键词： Graph embeddings

来源：评论

学校读者我要写书评

暂无评论

Spectroscopic photoacoustic denoising framework using hybrid analytical and data-free learning method

引用

Photoacoustics 2025年 44卷 100729页

作者： Lin, Fangzhou Gao, Shang Tang, Yichuan Ma, Xihan Murakami, Ryo Zhang, Ziming Obayemi, John D. Soboyejo, Winston O. Zhang, Haichong K. Department of Robotics Engineering Worcester Polytechnic Institute 100 Institute Road WorcesterMA01609 United States Department of Electrical & Computer Engineering Worcester Polytechnic Institute 100 Institute Road WorcesterMA01609 United States 60 Prescott Street WorcesterMA01605 United States Department of Mechanical & Materials Engineering Worcester Polytechnic Institute 100 Institute Road WorcesterMA01609 United States Department of Computer Science Worcester Polytechnic Institute 100 Institute Road WorcesterMA01609 United States

Spectroscopic photoacoustic (sPA) imaging uses multiple wavelengths to differentiate and quantify chromophores based on their unique optical absorption spectra. This technique has been widely applied in areas such as vascular mapping, tumor detection, and therapeutic monitoring. However, PA imaging is highly susceptible to noise, leading to a low signal-to-noise ratio (SNR) and compromised image quality. Furthermore, low SNR in spectral data adversely affects spectral unmixing outcomes, hindering accurate quantitative PA imaging. Traditional denoising techniques like frame averaging, though effective in improving SNR, can be impractical for dynamic imaging scenarios due to reduced frame rates. Advanced methods, including learning-based approaches and analytical algorithms, have demonstrated promise but often require extensive training data and parameter tuning. Moreover, spectral information preservation is unclear, limiting their adaptability for clinical usage. Additionally, training data is not always accessible for learning-based methods. In this work, we propose a Spectroscopic Photoacoustic Denoising (SPADE) framework using hybrid analytical and data-free learning method. This framework integrates a data-free learning-based method with an efficient BM3D-based analytical approach while preserving spectral integrity, providing noise reduction, and ensuring that functional information is maintained. The SPADE framework was validated through simulation, phantom, in vivo, and ex vivo studies. These studies demonstrated that SPADE improved image SNR by over 15 dB in high noise cases and preserved spectral information (R > 0.8), outperforming conventional methods, especially in low SNR conditions. SPADE presents a promising solution for preserving the accuracy of quantitative PA imaging in clinical applications where noise reduction and spectral preservation are critical. © 2025 The Authors

关键词： Image denoising

来源：评论

学校读者我要写书评

暂无评论

G-Mix: A Generalized Mixup Learning Framework Towards Flat Minima

IEEE Transactions on Artificial Intelligence

引用

IEEE Transactions on Artificial Intelligence 2025年

作者： Li, Xingyu Tang, Bo Mississippi State University Department of Electrical and Computer Engineering Mississippi StateMS39762 United States Worcester Polytechnic Instiute Department of Electrical and Computer Engineering WorcesterMA United States

Deep neural networks (DNNs) have demonstrated promising results in various complex tasks. However, such DNN models face challenges related to over-parameterization, particularly in scenarios where training data is scarce. In response to these challenges and to improve the generalization capabilities of DNNs, the Mixup technique has emerged, which effectively addresses the limitations posed by over-parameterization. Nevertheless, it still produces suboptimal outcomes. Inspired by the successful Sharpness-Aware Minimization (SAM) method, which establishes a connection between the sharpness of the training loss landscape and model generalization, we propose a new learning framework called Generalized-Mixup, which combines the strengths of Mixup and SAM for training DNN models. The theoretical analysis provided demonstrates how the developed G-Mix framework enhances generalization. Additionally, to further optimize DNN performance with the G-Mix framework, we introduce two novel algorithms: Binary G-Mix (BG-Mix) and Decomposed G-Mix (DG-Mix). These algorithms partition the training data into two subsets based on the sharpness-sensitivity of each example to address the issue of "manifold intrusion"in Mixup. Both theoretical explanations and experimental results reveal that the proposed BG-Mix and DG-Mix algorithms further enhance model generalization across multiple datasets and models, achieving state-of-the-art performance. © 2020 IEEE.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

Review of In-Pixel Gain Amplifiers and Readout Circuits for Low-Light CMOS Image Sensors

IEEE Sensors Reviews

引用

IEEE Sensors Reviews 2025年 2卷 95-111页

作者： Amr M. Maghraby Ibrahim Bozyel Islam T. Abougindia Suat U. Ay Electrical and Computer Engineering Department University of Idaho Moscow ID USA Electrical and Computer Engineering Department Worcester Polytechnique Institute Worcester MA USA Electronics Engineering Department Military Technical College Cairo Egypt

The low-light performance of image sensors can be enhanced by designing high-effective conversion gain signal paths from photon conversion sites to chip outputs. This can be achieved by using high-gain in-pixel amplifiers, high-gain column amplifiers, and by reducing the conversion capacitance of photon-sensing nodes. However, each of these approaches presents unique challenges and limitations that have restricted their widespread adoption in low-light applications. This article reviews the use of in-pixel gain amplifiers and their signal chain electronics in high-conversion-gain complementary metal–oxide–semiconductor (CMOS) image sensors over the past two decades. In-pixel gain amplifiers are classified into different categories according to the type of amplification technique used. Analyses of the column-referred conversion gain and the noise of each topology are presented alongside the different metrics used to characterize CMOS image sensor pixels for low-light imaging applications. The performance metrics of various in-pixel gain amplifiers are compared, providing a framework that highlights the best achieved input-referred noise in CMOS image sensors over the past 15 years. Furthermore, different tradeoffs are examined between optimizing conversion gain, pixel full-well capacity, and input read noise in both voltage and charge domains.

关键词： Noise Voltage Transistors Topology CMOS image sensors Reviews Imaging Capacitance Transconductors Circuits

来源：评论

学校读者我要写书评

暂无评论

Text-dependent Speaker Verification Challenge 2024: Exploring Shared and User-defined Passphrases

Text-dependent Speaker Verification Challenge 2024: Explorin...

引用

2025 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2025

作者： Zeinali, Hossein Lee, Kong Aik Alam, Jahangir Burget, Lukáš Department of Computer Engineering Amirkabir University of Technology Tehran Iran Department of Electrical and Electronic Engineering The Hong Kong Polytechnic University Kowloon Hong Kong Department of Computer Graphics and Multimedia Brno University of Technology Brno Czech Republic Computer Research Institute of Montreal Montreal Canada

ISBN: (纸本)9798350368741

In contrast to text-independent speaker verification, which has received significant attention from researchers and has many competitions dedicated to it, text-dependent speaker verification (TdSV) has been less explored recently. The TdSV Challenge 2024 was organized to analyze and explore novel methods for this type of speaker verification and aims to motivate participants to develop new approaches to TdSV, conduct comprehensive analyses, and investigate advanced techniques such as self-supervised learning. This challenge builds on the achievements of the short-duration speaker verification (SdSV) Challenges held in 2020 and 2021 and focuses specifically on TdSV in two distinct scenarios. The first scenario involves conventional TdSV, while the second focuses on speaker enrollment using user-defined passphrases. This paper provides a detailed description of both tasks, introduces the evaluation rules, and presents a comprehensive analysis of the results obtained from this challenge. © 2025 IEEE.

关键词： DeepMine dataset TdSV Challenge Text-dependent Speaker Verification User-defined passphrase

来源：评论

学校读者我要写书评

暂无评论

5DGWO-GAN:A Novel Five-Dimensional Gray Wolf Optimizer for Generative Adversarial Network-Enabled Intrusion Detection in IoT Systems

引用

computers, Materials & Continua 2025年第1期82卷 881-911页

作者： Sarvenaz Sadat Khatami Mehrdad Shoeibi Anita Ershadi Oskouei Diego Martín Maral Keramat Dashliboroun Department of Data Science Engineering University of HoustonHoustonTX 77204USA The WPI Business School Worcester Polytechnic InstituteWorcesterMA 01609USA School of Systems and Enterprises Stevens Institute of TechnologyHobokenNJ 07030USA Department of Computer Science Escuela de Ingeniería Informática de SegoviaUniversidad de ValladolidSegovia40005Spain Ernest G.Welch School of Art&Design Georgia State UniversityAtlantaGA 30303USA

The Internet of Things(IoT)is integral to modern infrastructure,enabling connectivity among a wide range of devices from home automation to industrial control *** the exponential increase in data generated by these interconnected devices,robust anomaly detection mechanisms are *** detection in this dynamic environment necessitates methods that can accurately distinguish between normal and anomalous behavior by learning intricate *** paper presents a novel approach utilizing generative adversarial networks(GANs)for anomaly detection in IoT ***,optimizing GANs involves tuning hyper-parameters such as learning rate,batch size,and optimization algorithms,which can be challenging due to the non-convex nature of GAN loss *** address this,we propose a five-dimensional Gray wolf optimizer(5DGWO)to optimize GAN *** 5DGWO introduces two new types of wolves:gamma(γ)for improved exploitation and convergence,and theta(θ)for enhanced exploration and escaping local *** proposed system framework comprises four key stages:1)preprocessing,2)generative model training,3)autoencoder(AE)training,and 4)predictive model *** generative models are utilized to assist the AE training,and the final predictive models(including convolutional neural network(CNN),deep belief network(DBN),recurrent neural network(RNN),random forest(RF),and extreme gradient boosting(XGBoost))are trained using the generated data and AE-encoded *** evaluated the system on three benchmark datasets:NSL-KDD,UNSW-NB15,and *** conducted on diverse IoT datasets show that our method outperforms existing anomaly detection strategies and significantly reduces false *** 5DGWO-GAN-CNNAE exhibits superior performance in various metrics,including accuracy,recall,precision,root mean square error(RMSE),and convergence *** proposed 5DGWO-GAN-CNNAE achieved the lowest RMSE values across the NSL-KDD,UNSW-NB15,and

关键词： Internet of things intrusion detection generative adversarial networks five-dimensional binary gray wolf optimizer deep learning

来源：评论

学校读者我要写书评

暂无评论

Single-Stage 48V-to-1V Regulator with a Half-Turn Transformer and Current-Doubler Rectifier 14

Single-Stage 48V-to-1V Regulator with a Half-Turn Transforme...

引用

14th Annual IEEE Applied Power Electronics Conference and Exposition, APEC 2025

作者： Xu, Xinmiao Li, Qiang Virginia Polytechnic Institute and State University Center for Power Electronics Systems Department of Electrical and Computer Engineering BlacksburgVA24061 United States

ISBN: (纸本)9798331516116

Single-stage 48V-to-1V regulated solutions for high-performance processors become increasingly popular. This paper presents a full-bridge transformer-based buck converter with the current-doubler rectifier. The transformer adopts a half-turn winding structure for reduced resistance and leakage inductance. Operating principles of the converter with the half-turn transformer and the converter's optimized printed circuit board layout are illustrated. Synchronous rectifiers are integrated in high-frequency loops and are treated carefully to reduce parasitic-inductor-induced loss. With the help of recent advances in gallium nitride (GaN) and silicon transistors, a 48V-to-1.8V GaN-based prototype switching at 700 kHz is constructed, which achieves 95.34% peak efficiency, 92.79% full-load efficiency (including gating loss) and 576 W/in3 power density with 50-A current per phase. The same hardware achieves 94.00% peak efficiency and 89.75% full-load efficiency with 55-A current per phase in 48V-to-1.0V conversion. Experimental results with 54-V input are also included in this paper, demonstrating the great potential of the single-stage solution for future data centers. © 2025 IEEE.

关键词： Electric rectifiers

来源：评论

学校读者我要写书评

暂无评论

The Utilization of a Parallel Computing Algorithm for Accelerating Switching-Level Modeling of Power Electronics Simulations in a T-Type PV Inverter 14

The Utilization of a Parallel Computing Algorithm for Accele...

引用

14th Annual IEEE Applied Power Electronics Conference and Exposition, APEC 2025

作者： Brown, Buck F. Wang, Liwei Zhang, Zheyu Enslin, Johan Li, Yi Clemson University Department of Electrical and Computer Engineering North CharlestonSC United States Clemson University School of Computing North CharlestonSC United States Rensselaer Polytechnic Institute Department of Electrical Computer and Systems Engineering TroyNY United States

ISBN: (纸本)9798331516116

High-fidelity, switching-level modeling of power electronics simulations can be computationally intensive and time-consuming. This computational burden has recently escalated further due to the increased number of converters implemented in a system, e.g., renewable energy systems. Unlike switching modeling, average modeling can alleviate the computational burden and adequately represent a converter's behavior for controller design. However, average modeling does not encapsulate switching harmonics, which are crucial for power electronics design, operation, and reliability studies. This paper proposes a novel way to accelerate the switching-level modeling of simulations by leveraging parallel computing techniques and showcases this technique in a specific case study. This new technique is called the average-to-switching (A2S) model. First, the basic concept of the A2S model is outlined. Second, a case study is designed for a two-stage, T-type PV inverter. Third, electrical and thermal waveforms produced by the case study are given and discussed. The results of testing the A2S model showcase a speedup of greater than 22 times the simulation benchmark while maintaining an R2 value of greater than 0.99. © 2025 IEEE.

关键词： Power electronics

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：