检索结果-内蒙古大学图书馆

InGaN micro-light-emitting diodes monolithically grown on Si:achieving ultra-stable operation through polarization and strain engineering

引用

Light(science & Applications) 2022年第10期11卷 2540-2548页

作者： Yuanpeng Wu Yixin Xiao Ishtiaque Navid Kai Sun Yakshita Malhotra Ping Wang Ding Wang Yuanxiang Xu Ayush Pandey Maddaka Reddeppa Walter Shin Jiangnan Liu Jungwook Min Zetian Mi Department of Electrical Engineering and Computer Science University of MichiganAnn ArborMI 48109USA Department of Materials Science and Engineering University of MichiganAnn ArborMI 48109USA

Micro or submicron scale light-emitting diodes(μLEDs)have been extensively studied recently as the next-generation display *** is desired that μLEDs exhibit high stability and efficiency,submicron pixel size,and potential monolithic integration with Si-based complementary metal-oxide-semiconductor(CMOS)*** such uLEDs,however,has remained a daunting *** polar nature of Ill-nitrides causes severe wavelength/color instability with varying carrier concentrations in the active *** etching-induced surface damages and poor material quality of high indium composition InGaN quantum wells(QWs)severely deteriorate the performance of uLEDs,particularly those emitting in the green/red *** we report,for the first time,μLEDs grown directly on Si with submicron lateral *** μLEDs feature ultra-stable,bright green emission with negligible quantum-confined Stark effect(QCSE).Detailed elemental mapping and numerical calculations show that the QCSE is screened by introducing polarization doping in the active region,which consists of InGaN/AlGaN QWs surrounded by an AlGaN/GaN shell with a negative Al composition gradient along the *** comparison with conventional GaN barriers,AlGaN barriers are shown to effectively compensate for the tensile strain within the active region,which significantly reduces the strain distribution and results in enhanced indium incorporation without compromising the material *** study provides new insights and a viable path for the design,fabrication,and integration of high-performance μLEDs on Si for a broad range of applications in on-chip optical communication and emerging augmented reality/mixed reality devices,and so on.

关键词： polarization diodes strain

来源：评论

学校读者我要写书评

暂无评论

LG-FAL: Locality-customized GSA Federated Active Learning 15

LG-FAL: Locality-customized GSA Federated Active Learning

引用

15th IEEE International Conference on Knowledge Graph, ICKG 2024

作者： Wang, Shuwen Zhu, Xingquan Portland State University Department of Computer Science Portland United States Florida Atlantic University Dept. of Electrical Engineering & Computer Science Boca Raton United States

ISBN: (纸本)9798331508821

Federated Learning (FL) is a learning paradigm which constructs machine learning models using decentralized datasets, with the goal of training a globally optimized model while preserving data privacy. In practice, annotating training data for FL is expensive and time-consuming. To solve this problem, researchers proposed to integrate active Learning (AL) strategy as Federated Active Learning (FAL) framework. However, existing research on FAL has two main limitations: (1) In the active learning part, more attention is paid to the global model, while ignoring the local information;(2) In the federated learning part, the popular Federated averaging (FedAvg) method relies on the assumption that the corresponding nodes in distributed neural networks share the same importance when averaging. In this paper, in order to tackle the two limitations, we propose a locality-customized GSA federated active learning (LG-FAL) method. Specifically, (i) both local and global information are taken into consideration to evaluate and select informative data samples to annotate for active learning;(ii) we integrate Gravitational Search Algorithm (GSA) to dynamically average local network parameters into the global parameter, by mimicking the principles of gravity and motion in the universe, for effective federated learning. As a result, LG-FAL can select a small subset of informative samples considering both local and global information, and at the same time, provide an improved trade-off between communication cost and learning accuracy. Experimental results show that LG-FAL significantly outperforms the current state-of-the-art baselines in terms of performance and effectiveness. © 2024 IEEE.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Public-Key Pseudoentanglement and the Hardness of Learning Ground State Entanglement Structure 39

Public-Key Pseudoentanglement and the Hardness of Learning G...

引用

39th Computational Complexity Conference, CCC 2024

作者： Bouland, Adam Fefferman, Bill Ghosh, Soumik Metger, Tony Vazirani, Umesh Zhang, Chenyi Zhou, Zixin Department of Computer Science Stanford University CA United States Department of Computer Science University of Chicago IL United States ETH Zürich Switzerland Department of Electrical Engineering and Computer Sciences University of California BerkeleyCA United States

ISBN: (纸本)9783959773317

Given a local Hamiltonian, how difficult is it to determine the entanglement structure of its ground state? We show that this problem is computationally intractable even if one is only trying to decide if the ground state is volume-law vs near area-law entangled. We prove this by constructing strong forms of pseudoentanglement in a public-key setting, where the circuits used to prepare the states are public knowledge. In particular, we construct two families of quantum circuits which produce volume-law vs near area-law entangled states, but nonetheless the classical descriptions of the circuits are indistinguishable under the Learning with Errors (LWE) assumption. Indistinguishability of the circuits then allows us to translate our construction to Hamiltonians. Our work opens new directions in Hamiltonian complexity, for example whether it is difficult to learn certain phases of matter. © Adam Bouland, Bill Fefferman, Soumik Ghosh, Tony Metger, Umesh Vazirani, Chenyi Zhang, and Zixin Zhou.

关键词： Ground state

来源：评论

学校读者我要写书评

暂无评论

Efficient Meeting Insights: NLP-Enhanced Summarization of Voice and Text 15

Efficient Meeting Insights: NLP-Enhanced Summarization of Vo...

引用

15th International Conference on Computing Communication and Networking Technologies, ICCCNT 2024

作者： Brahmachari, K. Sankar Hanumesh, Mangaluru Anitha, G. Pattanaik, Balachandra Amrita Vishwa Vidyapeetham Amrita School of Computing Department of Computer Science and Engineering Tamil Nadu Chennai India Wollega University School of Electrical Computer Engineering Nekemte Ethiopia

ISBN: (纸本)9798350370249

In order to improve meeting efficiency in the ever-changing workplace, research has mostly focused on sophisticated text summarization techniques such as machine learning, hybrid, semantic, and graph-based methods. These approaches have shown efficient in extracting key information from lengthy talks, hence dramatically improving decision-making and knowledge retention. However, their potential has been somewhat limited by a major concentration on processing material in a single language, which ignores the complex, multilingual environment of modern global enterprises. Our most recent research project tries to overcome this gap by combining these various summarizing approaches into acomplete, language-neutral ***, to address the specific difficulty of long meetings, when extracting insights might be especially difficult for those who are not there, we propose using the Longformer encoder-decoder paradigm. This methodology excels at distilling lengthy talks into digestible summaries, regardless of the original language, ensuring that critical insights are easily accessible to anybody, hence increasing inclusion and comprehension across multinational teams. © 2024 IEEE.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Performance-Guaranteed Finite-Time Tracking of Strict-Feedback Systems with Unknown Control Directions: A Novel Switching Mechanism

引用

IEEE Transactions on Automatic Control 2025年第6期70卷 4061-4068页

作者： Mao, Bing Wu, Xiaoqun Fan, Ziye Lu, Jinhu Chen, Guanrong Shenzhen University College of Computer Science and Software Engineering Shenzhen518060 China Wuhan University School of Mathematics and Statistics Hubei430072 China Beihang University School of Automation Science and Electrical Engineering Beijing100191 China City University of Hong Kong Department of Electrical Engineering Hong Kong Hong Kong

The problem of achieving performance-guaranteed finite-time exact tracking for uncertain strict-feedback nonlinear systems with unknown control directions is addressed. A novel logic switching mechanism with monitoring functions is designed to deal with unknown control directions. This approach is different from existing methods like the Nussbaum gain technique and smooth orientation functions, which are inapplicable to finite-time exact tracking. The main challenge lies in designing monitoring functions and determining switching rules, which is well resolved through three steps. Most significantly, in contrast to existing works that solely guarantee bounded tracking or asymptotic exact tracking, the new mechanism incorporates novel integral sliding mode techniques to achieve finite-time exact tracking with guaranteed performance. Finally, simulations are presented with comparison to demonstrate the effectiveness of the proposed methods. © 1963-2012 IEEE.

关键词： Feedback

来源：评论

学校读者我要写书评

暂无评论

Joint Device and Training Scheduling for Wireless Federated Learning

引用

IEEE Internet of Things Journal 2025年第12期12卷 18806-18819页

作者： Chen, Xiaobing Zhou, Xiangwei Zhang, Hongchao Sun, Mingxuan Zhao, Taibiao Louisiana State University Division of Electrical and Computer Engineering Baton RougeLA70803 United States Louisiana State University Department of Mathematics Baton RougeLA70803 United States Louisiana State University Division of Computer Science and Engineering Baton RougeLA70803 United States

The advent of ubiquitous computing devices in the Internet of Things (IoT) has resulted in an explosion of data. Traditional centralized machine learning models face challenges including limited bandwidth in wireless environments and privacy concerns due to their data aggregation approach. Federated learning addresses these challenges via decentralizing model training across numerous devices, leveraging model updates to enhance privacy and reduce communication overhead. To improve its cost efficiency, current research focuses on minimizing either time or energy costs but rarely both, and does not jointly optimize the parameters of device and training scheduling in the presence of system and data heterogeneity inherent in IoT networks. In our paper, we first introduce a multi-group transmission scheme and propose a comprehensive device scheduling framework, Group Scheduling on Orthogonal Frequency-Division Multiple Access (GS-OFDMA), to address time bottlenecks. Then we formulate a joint optimization problem for device and training scheduling that minimizes the total cost of training while ensuring model convergence. To tackle the resulting mixed integer nonlinear programming problem, we develop an iterative algorithm. Experimental results show that our approach significantly reduces the total cost by at least 35% across various real-world datasets and data distributions in comparison with random participant selection. The proposed GS-OFDMA protocol also exhibits higher time efficiency over other device scheduling schemes. © 2014 IEEE.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Intelligent wearable vision systems for the visually impaired in Saudi Arabia

引用

Neural Computing and Applications 2025年第14期37卷 8571-8591页

作者： Talaat, Fatma M. El-Shafai, Walid Soliman, Naglaa F. Algarni, Abeer D. El-Samie, Fathi E. Abd Faculty of Artificial Intelligence Kafrelsheikh University Kafrelsheikh33516 Egypt Faculty of Computer Science and Engineering New Mansoura University Gamasa35712 Egypt Security Engineering Lab Computer Science Department Prince Sultan University Riyadh11586 Saudi Arabia Department Electronics and Electrical Communications Engineering Faculty of Electronic Engineering Menoufia University Menouf32952 Egypt Department of Information Technology College of Computer and Information Sciences Princess Nourah Bint Abdulrahman University P.O. Box 84428 Riyadh11671 Saudi Arabia

Navigating the world with visual impairments presents unique challenges, often limiting independence and safety. This research introduces SafeStride, a novel algorithm designed to empower visually impaired individuals through real-time obstacle detection and navigation assistance. SafeStride functions through five interconnected phases: sensor data acquisition, data preprocessing, obstacle detection, path planning, and feedback and guidance. A combination of ultrasonic sensors, cameras, and inertial measurement units (IMUs) feed comprehensive environmental data into the system, enabling accurate obstacle detection and safe path planning. Evaluating SafeStride's performance involved the MNIST dataset for image recognition and a dedicated indoor object detection dataset. Additionally, comparisons were made against three deep learning models—convolutional neural network (CNN), long short-term memory (LSTM), and gated recurrent unit (GRU). The results showcased SafeStride's high accuracy and effectiveness in both obstacle detection and classification. Beyond its impressive performance, SafeStride represents a significant leap forward in navigation aids for the visually impaired. By offering a comprehensive solution that enhances safety and independence in diverse environments, it holds immense potential to improve the lives of countless individuals. Future efforts will focus on further optimizing the algorithm and testing its real-world capabilities, paving the way for a future where safe and independent navigation becomes a reality for all. All three deep learning models (CNN, LSTM, GRU) excelled in the experiment, achieving high accuracy above 0.98, strong recall and F1-scores near 0.99, and exceptional AUC-ROC scores exceeding 0.9998. © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2025.

关键词： Motion planning

来源：评论

学校读者我要写书评

暂无评论

Video and image quality improvement using an enhanced optimized dehazing technique

引用

Multimedia Tools and Applications 2024年 1-19页

作者： Ayoub, Abeer El-Shafai, Walid El-Samie, Fathi E. Abd Hamad, Ehab K. I. EL-Rabaie, El-Sayed M. Electrical Engineering Department Faculty of Engineering Aswan University Aswan81542 Egypt Department of Electronics and Electrical Communication Engineering Faculty of Electronic Engineering Menoufia University Menouf32952 Egypt Security Engineering Lab Computer Science Department Prince Sultan University Riyadh11586 Saudi Arabia Department of Information Technology College of Computer and Information Sciences Princess Nourah Bint Abdulrahman University P.O. Box 84428 Riyadh11671 Saudi Arabia

The detrimental effects of atmospheric haze frequently plague outdoor imagery. This phenomenon arises from the scattering of light by minute particles within the ambient environment surrounding the scene to be imaged. Haze engenders an overall whitening of the image, leading to diminished contrast. To address these issues and enhance the quality of hazy images and videos, an enhanced dehazing technique is proposed. The proposed technique includes image enhancement before the optimal dehazing process. The enhancement stage entails the implementation of both homomorphic processing and Contrast Limited Adaptive Histogram Equalization (CLAHE), serving to control the image dynamic range, while concurrently heightening the image contrast. The culminating stage encompasses an optimized dehazing technique, adept at expunging haze-induced artifacts from images. The homomorphic processing and CLAHE, applied in the pre-processing step, establish a foundation for the subsequent dehazing procedure. This proposed methodology is systematically applied to a gamut of visual outputs, including visible videos, Near-Infrared (NIR) frames, and authentic hazy images. Comparative evaluations of the proposed technique, homomorphic-processing-based enhanced dehazing, and standalone dehazing techniques is undertaken on different video types encompassing five frames. It is evident that the proposed technique, synergizing the homomorphic processing, CLAHE, and dehazing, outperforms alternative strategies. Furthermore, the proposed technique is subjected to a comparative study with various existing dehazing techniques over real hazy images. The assessment depends on Peak Signal-to-Noise Ratio (PSNR), correlation, and entropy metrics. The results underscore the efficacy of the proposed technique, particularly in terms of spectral entropy enhancement of dehazed frames. For both visible and NIR frames, the percentages of enhancement by the proposed technique are 17.66% and 118.48%, respectively. ©

关键词： Image enhancement

来源：评论

学校读者我要写书评

暂无评论

A Case Study on CALVEASE: AI Driven IoT Framework for Dystocia Prediction

A Case Study on CALVEASE: AI Driven IoT Framework for Dystoc...

引用

2023 IEEE Global Conference on Information Technologies and Communications, GCITC 2023

作者： Sharma, Anmol Hooda, Nishtha Hu, Yu Hen Singh Rana, Prashant University of Wisconsin-Madison Electrical and Computer Engineering Department Madison United States Indian Institute of Information Technology Department of Computer Science Una India Thapar Institute of Engineering and Technology Department of Computer Science and Engineering Patiala India

ISBN: (纸本)9798350308167

The ability to predict cow calving easiness cost-effectively, especially in the dairy industry where cattle suffer from a variety of unpredictable deadly illnesses and high breeding expenses assist farmers in improving reproductive performance and, thereby, farmers profitability. In this research work, machine learning and Internet-of-Things (IOT) based CALVEASE framework for monitoring dairy herd and predicting calving easiness is studied as a multi-classification problem. The CALVEASE framework collects activity information through sensors deployed in cow's collar belt in the first stage. The collected information stored at the cloud platform is utilized to train machine learning models for predicting the cow calving easiness in the second stage. On empirically testing the various models during field testing, random forest outperforms the other models showing an accuracy of 99.58% hence, the tested framework was found to be efficacious for boosting the dairy herd productivity. © 2023 IEEE.

关键词： Calving ease score Data analytics Machine learning Multi-class Classification

来源：评论

学校读者我要写书评

暂无评论

Adversarial Encoder-Driven Filter for Targeted Image Tagging: A Novel Approach to Visual Content Manipulation 6

Adversarial Encoder-Driven Filter for Targeted Image Tagging...

引用

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

作者： Mckee, Cole Flowers, Dominic Wood, Jesse Shafer, Ethan United States Military Academy Department of Electrical Engineering and Computer Science West PointNY10996 United States

ISBN: (纸本)9798331506520

computer vision, driven by artificial intelligence, has become pervasive in diverse applications such as self-driving cars and law enforcement. However, the susceptibility of these systems to attacks has raised significant concerns among researchers. This paper addresses the vulnerability of image tagging algorithms, particularly focusing on misclassifications induced by autoencoders. We present experiments conducted on Amazon Rekognition, where we developed a specialized autoencoder to manipulate the latent space, forcing it to align with specific tags. By integrating this manipulated latent space with other images, we demonstrate the ability to increase the confidence of a specific tag on Amazon Rekognition, leading to more false positives of the chosen tag. Our study showcases a practical method to exploit Amazon's Rekognition image tagging algorithm using a black box approach. © 2025 IEEE.

关键词： Adversarial machine learning

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：