检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Sharma, Manish Heard, Jamison Saber, Eli Markopoulos, Panos P. The Chester F. Carlson Center for Imaging Science Rochester Institute of Technology RochesterNY14623 United States The Department of Electrical and Microelectronic Engineering Rochester Institute of Technology RochesterNY14623 United States The Department of Electrical and Microelectronic Engineering The Chester F. Carlson Center for Imaging Science Rochester Institute of Technology RochesterNY14623 United States The Department of Electrical and Computer Engineering Department of Computer Science The University of Texas at San Antonio San AntonioTX78204 United States

While Convolutional Neural Networks (CNNs) excel at learning complex latent-space representations, their overparameterization can lead to overfitting and reduced performance, particularly with limited data. This, alongside their high computational and memory demands, limits the applicability of CNNs for edge deployment. Low-rank matrix approximation has emerged as a promising approach to reduce CNN parameters, but its application presents challenges including rank selection and performance loss. To address these issues, we propose an efficient training method for CNN compression via dynamic parameter rank pruning. Our approach integrates efficient matrix factorization and novel regularization techniques, forming a robust framework for dynamic rank reduction and model compression. We use Singular Value Decomposition (SVD) to model low-rank convolutional filters and dense weight matrices and we achieve model compression by training the SVD factors with back-propagation in an end-to-end way. We evaluate our method on an array of modern CNNs, including ResNet-18, ResNet-20, and ResNet-32, and datasets like CIFAR-10, CIFAR-100, and ImageNet (2012), showcasing its applicability in computer vision. Our experiments show that the proposed method can yield substantial storage savings while maintaining or even enhancing classification performance. Impact Statement—The rapid proliferation of edge computing and Internet of Things devices demands lightweight yet efficient machine learning models. Current Convolutional Neural Networks (CNNs), while powerful, are often too resource-intensive for these applications. Our proposed Dynamic Parameter Rank Pruning method automates the compression of CNNs during training without compromising performance. Specifically, our approach dynamically adapts the rank of layers during training based on data and task complexity, thereby eliminating the need for meticulous rank selection and manual adjustments either before or after training. This re

关键词： Factorization

来源：评论

学校读者我要写书评

暂无评论

High-Power Operation of 376 nm Laser with GaN Waveguide

High-Power Operation of 376 nm Laser with GaN Waveguide

引用

IEEE International Semiconductor Laser Conference

作者： Cheng Liu Qinchen Lin Guangying Wang Surjava Sanyal Swarnav Mukhopadhyay Shuqi Zhang Matthew Dwyer Matthew Seitz Tom Earles Nelson Tansu Jing Zhang Shubhra S. Pasayat Chirag Gupta Luke Mawst Department of Electrical and Computer Engineering University of Wisconsin-Madison Madison WI USA DRS Daylight Solutions Madison USA Department of Electrical and Microelectronic Engineering Rochester Institute of Technology Rochester NY USA School of Electrical and Electronic Engineering The University of Adelaide Adelaide Australia

Over 3 W output power from a 376 nm laser incorporating GaN as a waveguide layer was demonstrated under pulse operation. Comparisons between cleaved and etched facet devices are presented.

ISBN: (数字)9798350372991

ISBN: (纸本)9798350373004

Over 3 W output power from a 376 nm laser incorporating GaN as a waveguide layer was demonstrated under pulse operation. Comparisons between cleaved and etched facet devices are presented.

关键词： Semiconductor lasers Power lasers Waveguide lasers Semiconductor waveguides Power generation

来源：评论

学校读者我要写书评

暂无评论

Cross-Layer Optimizations for Ferroelectric Neuromorphic Computing

Cross-Layer Optimizations for Ferroelectric Neuromorphic Com...

引用

Midwest Symposium on Circuits and Systems (MWSCAS)

作者： A N M Nafiul Islam Kai Ni Abhronil Sengupta School of Electrical Engineering and Computer Science The Pennsylvania State University University Park PA USA Electrical & Microelectronic Engineering Department Rochester Institute of Technology Rochester NY USA

Hardware paradigms for neuromorphic computing mimic the functionalities of brain primitives in order to replicate similar area and energy advantages of biological nervous systems. Discovery of ferroelectricity in doped Hafnia has thrust Ferroelectronics, specifically Ferroelectric Field-effect transistors (FeFETs), at the forefront of realizing such device platforms for future data-centric applications. Despite the utility afforded by its intrinsic properties, e.g. CMOS-compatibility and scalability, harnessing their full potential requires a cross-layer design approach combining devices, circuits, and algorithms. In this paper, we review the recent developments looking at FeFETs and their application to enable low-power on-chip learning. We outline the unique opportunities emerging from device characterization and modelling results, that ultimately translate in novel algorithms and system-level benefits.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Mix-and-Conquer: Beamforming Design with Interconnected RIS for Multi-User Networks

Mix-and-Conquer: Beamforming Design with Interconnected RIS ...

引用

IEEE International Conference on Communications (ICC)

作者： Sajjad Nassirpour Naoki Kusashima Jose Flordelis Alireza Vahid Department of Electrical and Computer Engineering San Diego State University San Diego USA Lund laboratory Sony Europe B. V. Lund Sweden Department of Electrical and Microelectronic Engineering Rochester Institute of Technology Rochester USA

ISBN: (数字)9781728190549

ISBN: (纸本)9781728190556

We propose a new reconfigurable intelligent surface (RIS) structure, referred to as interconnected RIS (I-RIS), which allows the RIS elements to be interconnected and share the inci-dent signals using simple binary radio frequency (RF) switches and mix them into the reflecting signals. This structure enables multi-user scaling and requires fewer elements (i.e., a compact structure) compared to standard RIS (S-RIS), which assumes no interconnection between the elements. The I-RIS compact design makes it practical for deployment on space-limited nodes, e.g., unmanned aerial vehicles (UAVs). Hence, in this work, we propose a beamforming design based on I-RIS in a multi-user network, where we use binary RF switches as RIS elements. We show that our switch-based I-RIS offers a higher gain compared to an S-RIS using phase shifters. Finally, we introduce two optimization methods, sigmoid filled function (SFF) and semi-definite binary optimization (SBO), to optimize the RIS elements and evaluate their performance in terms of sum-rate and comolexity.

关键词： Radio frequency Array signal processing Phase shifters Optimization methods Switches Reconfigurable intelligent surfaces Autonomous aerial vehicles

来源：评论

学校读者我要写书评

暂无评论

Mix-and-Conquer: Beamforming Design with Interconnected RIS for Multi-User Networks

arXiv

引用

arXiv 2024年

作者： Nassirpour, Sajjad Kusashima, Naoki Flordelis, Jose Vahid, Alireza Department of Electrical and Computer Engineering San Diego State University San Diego United States Lund laboratory Sony Europe B.V. Lund Sweden Department of Electrical and Microelectronic Engineering Rochester Institute of Technology Rochester United States

We propose a new reconfigurable intelligent surface (RIS) structure, referred to as interconnected RIS (I-RIS), which allows the RIS elements to be interconnected and share the incident signals using simple binary radio frequency (RF) switches and mix them into the reflecting signals. This structure enables multi-user scaling and requires fewer elements (i.e., a compact structure) compared to standard RIS (S-RIS), which assumes no interconnection between the elements. The I-RIS compact design makes it practical for deployment on space-limited nodes, e.g., unmanned aerial vehicles (UAVs). Hence, in this work, we propose a beamforming design based on I-RIS in a multi-user network, where we use binary RF switches as RIS elements. We show that our switch-based I-RIS offers a higher gain compared to an S-RIS using phase shifters. Finally, we introduce two optimization methods, sigmoid filled function (SFF) and semi-definite binary optimization (SBO), to optimize the RIS elements and evaluate their performance in terms of sum-rate and complexity. Copyright © 2024, The Authors. All rights reserved.

关键词： Beamforming

来源：评论

学校读者我要写书评

暂无评论

Modified Kelvin-Voigt fractional derivative model for viscoelasticity measurement in optical coherence elastography

引用

Chinese Optics Letters 2025年第1期23卷 93-101页

作者： Chenming Yang Zhongliang Li Nan Nan Teng Liu Yaoli Luo Xiangzhao Wang Advanced Light Source and System R&D Center Department of Advanced Optical and Microelectronic EquipmentShanghai Institute of Optics and Fine MechanicsChinese Academy of SciencesShanghai 201800China Center of Materials Science and Optoelectronics Engineering University of Chinese Academy of SciencesBeijing 100049China School of Optical-Electrical and Computer Engineering University of Shanghai for Science and TechnologyShanghai 200093China State Key Laboratory of Extreme Photonics and Instrumentation College of Optical Science and EngineeringZhejiang UniversityHangzhou 310027China

Optical coherence elastography(OCE)can quantitatively obtain the viscoelasticity of tissues using rheological models and is widely applied to the clinical diagnosis of ***,commonly used rheological models in OCE do not account for the distinctive dependence of high-frequency storage and loss moduli on frequency in tissues,which results in the rheological models failing to accurately measure the viscoelastic properties of *** this paper,a modified Kelvin-Voigt fractional derivative model is presented based on the power-law behavior of soft tissues and the dependence of high-frequency complex shear modulus on frequency in living *** the rheometer and OCE tests,the modified model can provide the prediction of the power-law relationship between the low-frequency shear viscosity and frequency;compared with the Kelvin-Voigt and Kelvin-Voigt fractional derivative models,the modified model has a higher goodness-of-fit(accuracy>96%)for the high-frequency storage moduli of gelatin ***,the proposed model can reduce the root mean square error of fit by approximately 83%for the high-frequency(1–128 kHz)storage modulus of the polydimethylsiloxane phantoms obtained from publicly available ***,the modified model accurately predicts the mechanical properties of biomimetic materials over a wide frequency range,with the potential to more accurately reflect pathological changes in tissues.

关键词： optical coherence tomography optical coherence elastography viscoelasticity rheological model

来源：评论

学校读者我要写书评

暂无评论

POLAR: Performance-aware On-device Learning Capable Programmable Processing-in-Memory Architecture for Low-Power ML Applications

POLAR: Performance-aware On-device Learning Capable Programm...

引用

Euromicro Symposium on Digital System Design

作者： Sathwika Bavikadi Purab Ranjan Sutradhar Mark A. Indovina Amlan Ganguly Sai Manoj Pudukotai Dinakarrao Department of Electrical and Computer Engineering George Mason University Fairfax USA Department of Computer Engineering Rochester Institute of Technology Rochester USA Department of Electrical and Microelectronic Engineering Rochester Institute of Technology Rochester USA

ISBN: (数字)9781665474047

ISBN: (纸本)9781665474054

Improving the performance of real-time Traffic Sign Recognition (TSR) applications using Deep Learning (DL) algorithms such as Convolutional Neural Networks (CNN) on software platforms is challenging due to the sheer computational complexity of these algorithms. In this work, we adopt a hardware-software combined approach to address this issue. We introduce a data-centric Processing-in-Memory (PIM) architecture that leverages Look-up-Table (LUT)-based processing for minimal data movement and superior performance and efficiency. Despite the superior performance, the limited available memory in PIM makes it complex to deploy deep CNNs. We propose merging CNN layers in this work to meet the limited resource constraints. One specific challenge in the TSR is the continuous change in the deployed environment, which makes a CNN model train over static data, leading to performance degradation over time. To address these challenges, we introduce a lightweight, performance-aware Generative Adversarial Network (GAN)-based on-device learning on PIM architecture. This compact CNN on PIM architecture attains data-level parallelism and reduces pipelining delays and makes it easier for on-device training and inference. Evaluation is performed on multiple state-of-the-art DL networks such as LeNet, AlexNet, ResNet using the German Traffic Sign Recognition Benchmark (GTSRB) Dataset, and the Belgium Traffic Sign Dataset (BTSD). With the proposed learning technique, it is observed to achieve maximum accuracy of 92.8% and 89.27% on GTSRB, and BTSD datasets. Also, it is observed the proposed mechanism maintains an average accuracy to be above 85% despite changes in the environment on all the CNNs deployed on the PIM accelerator.

关键词： Training Software algorithms Merging Neural networks computer architecture Generative adversarial networks Software

来源：评论

学校读者我要写书评

暂无评论

Low-loss Cd-substituted Mg ferrites with matching impedance for high-frequency-range antennas

引用

Journal of Magnesium and Alloys 2021年第4期9卷 1403-1413页

作者： Gongwen Gan Dainan Zhang Jie Li Gang Wang Xin Huang Yan Yang Yiheng Rao Xueying Wang Huaiwu Zhang Ray T.Chen State Key Laboratory of Electronic Thin Films and Integrated Devices University of Electronic Science and Technology of ChinaNo.4Section 2North Jianshe RoadChengdu 610054China Microelectronic Research Center Department of Electrical and Computer EngineeringUniversity of Texas at AustinAustinTX 78758USA

The effects of Cd^(2+) ions on the microstructure,magnetic properties,and dielectric properties of Bi_(2)O_(3)-added MgFe_(2)O_(4) ferrites(Cd_(x)Mg_(1-x)Fe_(2)O_(4),x=0.00,0.15,0.30 and 0.45)are obtained by adopting the solid-state reaction method at a low temperature(910℃).The objective is to achieve matching impedances,low magnetic and dielectric losses(tanδμand tanδε,respectively),and a relatively large miniaturization factor to reduce antenna *** results indicate that the cations occupying the tetrahedral(A)and octahedral(B)ion sites are redistributed,resulting in an enhanced super-exchange interaction between the two *** a result,improved magnetization,including the increase in saturation magnetization(41.74 emu/g)and decrease in coercivity(63.75 Oe),is *** real part of permeability(μ')also increases with increasing concentration of Cd^(2+) *** x is 0.15,matching impedances with equivalent μ'and ε'values are obtained over a long frequency range(1–150MHz).Moreover,the formation of a dense microstructure guarantees that losses occur at low orders of magnitude(tanδμ≈10−2 and tanδε≈10−3).Accordingly,these properties afford wide application perspectives for the proposed compounds in the high-frequency region,i.e.,from high-frequency to very-high-frequency bands.

关键词： Cd-substituted Mg compounds Magnetic properties Dielectric properties Matching impedance Low loss High frequency antenna

来源：评论

学校读者我要写书评

暂无评论

Face recognition using pca implemented on raspberry pi 11th

Face recognition using pca implemented on raspberry pi

引用

11th National Technical Symposium on Unmanned System Technology, NUSYS 2019

作者： Mohammed, Ibrahim Majid Al-Dabagh, Mustafa Zuhaer Nayef Ahmad, Muhammad Imran Isa, Mohd Nazrin Md School of Computer and Communication Engineering University Malaysia Perlis ArauPerlis Malaysia School of Microelectronic Engineering University Malaysia Perlis ArauPerlis Malaysia Department of Computer Science College of Science Knowledge University ErbilKurdistan Region Iraq

ISBN: (纸本)9789811552809

Currently, many consumers electronic are utilizing personal identification technique such as ID, fingerprint, face, Iris and palmprint recognition for security reason. Among other biometric traits, face image is easy to capture using Pi Camera. This paper proposes facial recognition using principle component analysis (PCA) method implemented using Raspberry Pi embedded processor. The algorithm is model using PC based method to identify the best eigenvector to represent facial image. The pre-calculated statistical parameter is then used to implement the PCA algorithm in Raspberry Pi embedded processor. OpenCV image processing libraries is use to support the basic task of image pre-processing such as cropping, resize and color conversion. In the propose work, the whole system is implemented in low cost processor to evaluate the performance in terms of recognition rates and processing time. The proposed method includes three main phases. The pre-processing phase processes facial ORL image and to collect significant information. The second phase extracts important features from image and decreases the size of the image using the PCA method. This phase uses a linear projection technique to decrease redundancies and remove noise from the image. Moreover, this strategy also improves the strength of discrimination power in the feature space. The Euclidean distance classifier is utilized in the third phase of classification. The best recognition rates achieved using the propose method is 82.5% implemented using low cost embedded system. © Springer Nature Singapore Pte Ltd 2021.

关键词： Python

来源：评论

学校读者我要写书评

暂无评论

Design of novel circular lattice photonic crystal fiber suitable for transporting 48 OAM modes

引用

Optoelectronics Letters 2021年第8期17卷 501-506页

作者： Alaaeddine Rjeb Habib Fathallah Saleh Chebaane Mohsen Machhout Laboratory of Electronic and Microelectronic Physics DepartmentFaculty of Sciences of MonastirUniversity of MonastirMonastir 5000Tunisia Laboratory of Artificial Intelligence and Data Engineering Applications Computer DepartmentFaculty of Sciences of BizerteUniversity of CarthageBizerte 7000Tunisia Physics Department College of ScienceUniversity of Hai'ilKingdom of Saudi Arabia

In this paper,we propose and numerically investigate a novel circular lattice photonic crystal fiber(CL-PCF)using controllable GeO_(2) doped silica,suitable for modes carrying quantized orbital angular momentum(OAM).Large effective index separations between 25 supported vector modes(≥10^(-4))are confirmed over large bandwidth(C+L bands)leading to 48 OAM modes bearing *** simulations show that the modes in the proposed CLPCF have good features including low and flat dispersion(within 51.82 ps/km/nm),low confinement loss(lower than 0.002 d B/m),high effective mode area(88.5μm^(2))and low nonlinearity(1.22 W^(-1)·km^(-1)).These promising results show that the proposed CL-PCF could be an arguably candidate in fiber-based OAM multiplexing or other applications using OAM states.

关键词： fiber OAM circular

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：