检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Mei, Zhen Cai, Kui Shi, Long Li, Jun School of Electronic and Optical Engineering Nanjing University of Science and Technology Nanjing210094 China National Mobile Communi-cations Research Laboratory Southeast University Nanjing210096 China Science and Math Cluster Singapore University of Technology and Design 487372 Singapore

Due to the crossbar array architecture, the sneak-path problem severely degrades the data integrity in the resistive random access memory (ReRAM). In this letter, we investigate the channel quantizer design for ReRAM arrays with multiple reads, which is a typical technique to improve the data recovery performance of data storage systems. Starting with a quantized channel model of ReRAM with multiple reads, we first derive a general approach for designing the channel quantizer, for both single-bit and multiple-bit quantizations. We then focus on the single-bit quantization, which is highly suitable for practical applications of ReRAM. In particular, we propose a semi-analytical approach to design the multiple-read single-bit quantizer with less complexity. We also derive the theoretical bit-error probability of the optimal single-bit detector/quantization as the benchmark. Results indicate that the multiple-read operation is effective in improving the error rate performance of ReRAM. Moreover, our proposed multiple-read detector outperforms the prior art detector and achieves the performance of the optimal detector. Copyright © 2024, The Authors. All rights reserved.

关键词： Memory architecture

来源：评论

学校读者我要写书评

暂无评论

2D Points Curve Reconstruction Survey and Benchmark

2D Points Curve Reconstruction Survey and Benchmark

引用

43rd Annual Conference on European Association for Computer Graphics, EUROGRAPHICS 2022

作者： Ohrhallinger, S. Peethambaran, J. Parakkat, A.D. Dey, T.K. Muthuganapathy, R. Institute of Visual Computing and Human-Centered Technology TU Wien Austria Department of Math and Computing Science Saint Mary's University Halifax Canada Department of Computer Science and Engineering Indian Institute of Technology Guwahati India Department of Computer Science Purdue University IN United States Department of Engineering Design Indian Institute of Technology Madras India

ISBN: (纸本)9783038681724

Curve reconstruction from unstructured points in a plane is a fundamental problem with many applications that has generated research interest for decades. Involved aspects like handling open, sharp, multiple and non-manifold outlines, run-time and provability as well as potential extension to 3D for surface reconstruction have led to many different algorithms. We survey the literature on 2D curve reconstruction and then present an open-sourced benchmark for the experimental study. Our unprecedented evaluation of a selected set of planar curve reconstruction algorithms aims to give an overview of both quantitative analysis and qualitative aspects for helping users to select the right algorithm for specific problems in the field. Our benchmark framework is available online to permit reproducing the results and easy integration of new algorithms. © 2022 The Author(s) Eurographics Proceedings © 2022 The Eurographics Association.

关键词： Surveys

来源：评论

学校读者我要写书评

暂无评论

A Binary PSO Based Ensemble Under-Sampling Model for Rebalancing Imbalanced Training Data

arXiv

引用

arXiv 2025年

作者： Li, Jinyan Wu, Yaoyang Fong, Simon Tallón-Ballesteros, Antonio J. Yang, Xin-She Mohammed, Sabah Department of Computer and Information Science University of Macau Taipa China Department of Languages and Computer Systems University of Seville Seville Spain Design Engineering and Math School of Science & Technology Middlesex University London United Kingdom Department of Computer Science Lakehead University Thunder Bay Canada

Ensemble technique and Under-sampling technique are both effective tools for resolvig imbalanced dataset classification problems, which commonly denotes the quantitative imbalance of a binary class dataset where minority class is the target class. In this paper, a novel ensemble method absorbing the advantages of both ensemble learning for biasing classifiers and a new evolutional under-sampling method is proposed. The under-sampling method is named Binary PSO instance selection, it gathers with ensemble classifiers to find the most suitable length and combination of majority class samples to build a new dataset with minority class samples. The proposed method adopts multi-objective strategy, objectives of this method is to improve the performances of imbalanced classification and to guarantee the maximum integrity of the original dataset. We examined our proposed method, the Binary PSO instance selection by comparing its performance of processing imbalanced dataset with several other conventional basic ensemble methods. Experiments are also conducted with Binary PSO instance selection wrapping with ensemble classifiers for further improvement of imbalanced classification. Based on comaprison experiments, our proposed methods outperform single ensemble methods, state-of-the-art under-sampling methods as well as their combinations with traditional PSO instance selection algorithm. Copyright © 2025, The Authors. All rights reserved.

关键词： Data assimilation

来源：评论

学校读者我要写书评

暂无评论

Virtual Field Trip of the Central Heating and Chiller Plant as an Instructional Tool for Thermodynamics Education

Virtual Field Trip of the Central Heating and Chiller Plant ...

引用

Frontiers in Education (FIE) Conference

作者： Oai Ha Seth Berrier Keif Oss Dept. of Engineering and Technology University of Wisconsin-Stout Menomonie Wisconsin Dept. of Math Stats and Computer Science University of Wisconsin-Stout Menomonie Wisconsin Dept. of Design University of Wisconsin-Stout Menomonie Wisconsin

This work in progress introduces a framework for developing a virtual tour of the central heating and chiller plant (the Plant) at a university in the US Midwest for teaching and learning Thermodynamics at the campus. Traditionally, this theory-heavy and abstract course is taught using textbooks, lectures, problem solving, and some laboratory experiments. The majority of course content lacks the tangible connections between engineering devices and systems in a close real-world application, making the course challenging and uninteresting for many students. In this project, 360-degree panoramic photos and proper authoring software are employed to render an immersive virtual environment of the Plant that students can visit at any time and from anywhere. The virtual field trip would provide students with tangible insights into thermodynamic and heat transfer processes from over 25 main engineering equipment, devices, and systems at the Plant. Among them are boilers, chillers, a steam turbine, heat exchangers, cooling towers, blowers, and economizers. The follow-up quizzes and assessments are designed to contribute to the course learning outcomes, such as an understanding of phase transitions, thermodynamic properties of steam, and thermal efficiencies of engineering devices. The virtual immersive experiences can help students enhance their learning experiences and engagement in the course's materials without the costs and dangers inherent in a physical tour of the Plant facility. A preliminary result of the virtual tour's evaluation with a group of students and faculty is also included to provide inputs to the next phase of the project.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Giant rectification in segmented, strongly interacting spin chains despite the presence of perturbations

引用

Physical Review E 2021年第5期103卷 052143-052143页

作者： Kang Hao Lee Vinitha Balachandran Dario Poletti Science and Math Cluster Singapore University of Technology and Design 8 Somapah Road 487372 Singapore Engineering Product Development Pillar Singapore University of Technology and Design 8 Somapah Road 487372 Singapore

Balachandran et al. [Phys. Rev. Lett. 120, 200603 (2018)] presented a segmented XXZ spin chain with zero anisotropy in one half and a large anisotropy on the other half that gave rise to a spin current rectification which is perfect in the thermodynamic limit. Here we extend the previous study to segmented chains with interacting integrable as well as nonintegrable halves, considering even cases in which no ballistic transport can emerge in either half. We demonstrate that, also in this more general case, it is possible to obtain giant rectification when the two interacting half chains are sufficiently different. We also show that the mechanism causing this effect is the emergence of an energy gap in the excitation spectrum of the out-of-equilibrium insulating steady state in one of the two biases. Finally, we demonstrate that in the thermodynamic limit there is no perfect rectification when each of the two half chains is interacting.

关键词： Magnons Nonequilibrium statistical mechanics Open quantum systems Quantum transport Spin current 1-dimensional spin chains Quantum spin models Strongly correlated systems Exact diagonalization Many-body techniques

来源：评论

学校读者我要写书评

暂无评论

LLM Acceleration on FPGAs: A Comparative Study of Layer and Spatial Accelerators

LLM Acceleration on FPGAs: A Comparative Study of Layer and ...

引用

IEEE Central America and Panama Convention (CONCAPAN)

作者： Luis D. Prieto-Sibaja Luis G. Leόn-Vega Indra Leόn-Vega Jorge Castro-Godínez Stefano Cozzini Computer Engineering School Costa Rica Institute of Technology Cartago Costa Rica Math. Inf. and Geoscience Dep. University of Trieste Trieste Italy Industrial Design Engineering School Instituto Tecnolόgico de Costa Rica Cartago Costa Rica Electronics Engineering School Costa Rica Institute of Technology Cartago Costa Rica Institute of Research and Technology AREA Science Park Trieste Italy

ISBN: (数字)9798350366723

ISBN: (纸本)9798350366730

Large Language Models (LLMs) have emerged as the leading technique in Natural Language Processing due to their remarkable capabilities. These models are exceptionally complex, often utilising billions of parameters and consuming tens of gigabytes of memory unless optimised. Typically, LLMs are trained and accelerated using general-purpose Graphics Processing Units (GPUs), with a single inference potentially occupying an entire GPU. Consequently, cloud services may require hundreds of GPUs to deliver high-quality AI assistant services. This demand for extensive hardware acceleration allows exploring alternative architectures that offer improved execution time and energy *** study evaluates Field-Programmable Gate Arrays (FP-GAs), renowned for their flexibility and efficiency, to accelerate LLMs. We specifically compare the resource consumption and latency of two distinct architectures: layer and spatial accelerators. Analysing the Llama 2-7B model, we identified potential optimisation opportunities within its composition and operational graphs. Our most successful implementations demonstrate a performance improvement ranging from 1.37× to 10.98× over two AMD EPYC processors with 64 cores each. Moreover, our results indicate that, with further refinement, FPGAs have the potential to surpass GPU performance for LLM inference, showcasing their feasibility as a viable alternative for this demanding application.

关键词： Large language models High performance computing Graphics processing units Computer architecture Logic gates Distance measurement Natural language processing Field programmable gate arrays Hardware acceleration Optimization

来源：评论

学校读者我要写书评

暂无评论

Real-time impurity solver using Grassmann time-evolving matrix product operators

引用

Physical Review B 2024年第16期109卷 165113-165113页

作者： Ruofan Chen Xiansong Xu Chu Guo College of Physics and Electronic Engineering and Center for Computational Sciences Sichuan Normal University Chengdu 610068 China Science and Math Cluster Singapore University of Technology and Design 8 Somapah Road Singapore 487372 Key Laboratory of Low-Dimensional Quantum Structures and Quantum Control of Ministry of Education Department of Physics and Synergetic Innovation Center for Quantum Effects and Applications Hunan Normal University Changsha 410081 China

An emergent and promising tensor-network-based impurity solver is to represent the Feynman-Vernon influence functional as a matrix product state, where the bath is integrated out analytically. Here we present an approach to calculate the equilibrium impurity spectral function based on the recently proposed Grassmann time-evolving matrix product operators method. The central idea is to perform a quench from a separable impurity-bath initial state as in the nonequilibrium scenario. The retarded Green's function G(t+t0,t′+t0) is then calculated after an equilibration time t0 such that the impurity and bath are approximately in thermal equilibrium. There are two major advantages of this method. First, since we focus on real-time dynamics, we do not need to perform the numerically ill-posed analytic continuation as in imaginary- time evolution-based methods. Second, the required bond dimension of the matrix product state in real-time calculations is observed to be much smaller than that in imaginary-time calculations, leading to a significant improvement in numerical efficiency. The accuracy of this method is demonstrated using the single-orbital Anderson impurity model and benchmarked against the continuous-time quantum Monte Carlo method.

关键词： Open quantum systems Anderson impurity model Matrix product states

来源：评论

学校读者我要写书评

暂无评论

A Multi-scale Binarized Neural Network Application Based on All Programmable System on Chip

A Multi-scale Binarized Neural Network Application Based on ...

引用

IEEE International Symposium on Embedded Multicore Socs (MCSoC)

作者： Maoyang Xiang T. Hui Teo Singapore University of Technology and Design Singapore Singapore Science Math & Technology Cluster / Engineering Product Development Singapore University of Technology and Design Singapore Singapore

ISBN: (纸本)9781728187525

Binary neural networks (BNNs) are particularly well-suited for low-power embedded devices with limited computational capabilities. Due to the binary weight parameters, it significantly reduces memory footprint and arithmetic logic unit operations. Nevertheless, one of the disadvantages of BNN is low accuracy and sharp optimization space. Several studies of BNNs have recently shown improved accuracy in various tests via more operations and more complicated topologies. This approach, however, is incompatible with the embedded BNN application since it requires complicated data type translation. Hence, We propose a novel approach for the BNN application on the embedded system with multi-scale neural network topology in this research from two optimization perspectives: hardware structure and BNN topology, which preserves more low-level information during the feed-forward process with few operations. Our network topology achieves 91.3% accuracy for the CIFAR-10 dataset, one of the highest recorded by BNN and can process 537 tiny pictures per second when deployed on an All programmable System on Chip (APSoc) device with 4.4W power consumption.

关键词： Power demand Embedded systems Network topology Multicore processing Neural networks Hardware Topology

来源：评论

学校读者我要写书评

暂无评论

Broadband strong optical dichroism in topological Dirac semimetals with Fermi velocity anisotropy

引用

Chinese Physics B 2020年第7期29卷 128-134页

作者： J Lim K J A Ooi C Zhang L K Ang Yee Sin Ang Science Math and TechnologySingapore University of Technology and DesignSingapore School of Energy and Chemical Engineering Xiamen University MalaysiaSelangor Darul Ehsan 43900Malaysia College of Chemistry and Chemical Engineering Xiamen UniversityXiamen 361005China School of Physics University of WollongongNorthfields AvenueNew South Wales 2522Australia

Prototypical three-dimensional(3D)topological Dirac semimetals(DSMs),such as Cd3As2 and Na3Bi,contain electrons that obey a linear momentum-energy dispersion with different Fermi velocities along the three orthogonal momentum *** being extensively studied in recent years,the inherent Fermi velocity anisotropy has often been neglected in the theoretical and numerical studies of 3D *** this omission does not qualitatively alter the physics of light-driven massless quasiparticles in 3D DSMs,it does quantitatively change the optical coefficients which can lead to nontrivial implications in terms of nanophotonics and plasmonics *** we study the linear optical response of 3D DSMs for general Fermi velocity values along each *** the signature conductivity-frequency scaling,σ(ω)∝ω,of 3D Dirac fermion is well-protected from the Fermi velocity anisotropy,the linear optical response exhibits strong linear dichroism as captured by the universal extinction ratio scaling law,Λi j=(vi/v j)^2(where i=j denotes the three spatial coordinates x,y,z,and vi is the i-direction Fermi velocity),which is independent of frequency,temperature,doping,and carrier scattering *** Cd3As2 and Na3Bi3,an exceptionally strong extinction ratio larger than 15 and covering a broad terahertz window is *** findings shed new light on the role of Fermi velocity anisotropy in the optical response of Dirac semimetals and open up novel polarization-sensitive functionalities,such as photodetection and light modulation.

关键词： topological Dirac semimetal optical anisotropy linear dichroismn linear response theory

来源：评论

学校读者我要写书评

暂无评论

Non-Markovian effects in stochastic resonance in a two level system

arXiv

引用

arXiv 2022年

作者： Chen, Ruofan Xu, Xiansong College of Physics and Electronic Engineering Center for Computational Sciences Sichuan Normal University Chengdu610068 China Science and Math Cluster Singapore University of Technology and Design 8 Somapah Road Singapore487372 Singapore

Stochastic resonance is a phenomenon where the response signal to external driving is enhanced by environment noise. In quantum regime, the effect of environment is often intrinsically non-Markovian. Due to the combination of such non-Markovian quantum noise and external driving force, it is difficult to evaluate the correlation function and hence the power spectrum. Nevertheless, a recently developed algorithm, which is called time-evolving matrix product operators (TEMPO), and its extensions provide an efficient and numerically exact approach for this task. Using TEMPO we investigate non-Markovian effects in quantum stochastic resonance in a two level system. The periodic signal and the time-averaged asymptotic correlation function, along with the power spectrum, are calculated. From the power spectrum the signal-to-noise ratio is evaluated. It is shown that both signal strength and signal-to-noise ratio are enhanced by non-Markovian effects, which indicates the importance of non-Markovian effects in quantum stochastic resonance. In addition, we show that the non-Markovian effects can shift the peak position of the background noise power spectrum. Copyright © 2022, The Authors. All rights reserved.

关键词： Signal to noise ratio

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：