检索结果-内蒙古大学图书馆

2024 8th International Conference on Electrical, Mechanical and Computer Engineering (ICEMCE)

作者： Sheng Qin Yang Zuo Shengkai Jiang Yuekang Qian Heng Li Wehua Gu School of Microelectronics (School of Integrated Circuits) Nanjing University of Science and Technology Key Laboratory of Near-Range RF Sensing ICs & Microsystems (Nanjing University of Science and Technology) Ministry of Education Nanjing China

ISBN: (数字)9798331506230

ISBN: (纸本)9798331506247

The traditional optical power control technology of fiber optic gyroscope is to sample, demodulate and other processing of the electrical signal output by the PIN-FET to obtain real-time optical power, and then maintain the stability of optical power through feedback control. In this paper, the accurate optical power information is obtained in real time through the Micro-photodetector (MPD), and then the PID optical power control algorithm is introduced, with the digital current source as the controlled object, and the optical power fluctuation under full temperature conditions is minimized by adjusting the control parameters. At the same time, in order to cope with the problem of slow system start at the extreme temperature, the double-layer PID algorithm is used to nest, and the PID fast start algorithm is started when the fiber optic gyro system is just powered on, which makes the optical power approach the target faster and more conducive to fast start, and after the start is completed, it is switched to the optical power control algorithm to achieve accurate control, and the smooth switching of the two sets of PID algorithms makes the system have the advantages of fast startup and small optical power fluctuation at full temperature.

关键词： optical fibers Integrated optics optical feedback Power control Real-time systems Gyroscopes Temperature control optical devices Thermal stability optical control

来源：评论

学校读者我要写书评

暂无评论

Exploiting Functional Approximation on Decision-Tree Based Multiple Classifier systems

Exploiting Functional Approximation on Decision-Tree Based M...

引用

IFIP International Conference on Very Large Scale Integration (VLSI-SoC)

作者： Mario Barbareschi Salvatore Barone Antonio Emmanuele Nicola Mazzocca Department of Electrical Engineering and Information Technologies University of Naples Federico II Naples Italy

ISBN: (数字)9798331539672

ISBN: (纸本)9798331539689

Multiple Classifier systems (MCSs) have been in-creasingly designed to take advantage of hardware features, such as high parallelism and computational power, to guarantee higher throughput and lower latency. Although the combination of multiple classifiers leads to high classification accuracy, the required area overhead makes the design of a hardware accelerator unfeasible, hindering the adoption of commercial configurable devices. For this reason, in this paper, we exploit the Approximate Computing (AxC) design paradigm to automatically generate approximated hardware implementations of MCSs by trading hardware area overhead off for classification accuracy. In particular, we propose an algorithm that identifies the resiliency source of the model and uses it to introduce approximation with minimum accuracy loss.

关键词： Accuracy Correlation Computational modeling Very large scale integration Parallel processing Throughput Classification algorithms Proposals Hardware acceleration Resilience

来源：评论

学校读者我要写书评

暂无评论

An Investigation for Cursive Context-Specific Printed Script Recognition Techniques

An Investigation for Cursive Context-Specific Printed Script...

引用

IEEE SSD International Multi-Conference on systems, Signals and devices

作者： Humera Rafique Tariq Javid Dept. of Electrical Engineering Graduate School of Engineering Sciences & Information Technology Hamdard University Karachi Pakistan

Selecting suitable techniques and datasets for a machine learning problem is a great challenge. This research investigates the available machine learning architectures, datasets, and software tools, for cursive context-specific printed text. Despite the existence of millions of speakers around the globe and rich literary history of more than thousands of years, it is hard to find computational linguistic works related to the Punjabi Shahmukhi script, a member of Perso-Arabic context-specific script – a low-resource language family. An investigation of related work through collected statistics inspired by the popularity and success of artificial neural networks and associated techniques have presented. The historical development of the script follows it in a literary sense and the importance and motivation in the development of optical character recognition systems, algorithms, and techniques. The paper discussed the available datasets and the need for a custom dataset. Based on recent research trends in machine learning frameworks, the work presents the most popular deep learning techniques, convolutional neural networks, and recursive neural networks for cursive context-specific scripts. This paper incorporates actual knowledge to promote research in machine learning and natural language processing disciplines and the selection of algorithms, architectures, and resources.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Design-Space Exploration of Systolic Array for Edge Inferencing Applications 23

Design-Space Exploration of Systolic Array for Edge Inferenc...

引用

Proceedings of the Third International Conference on AI-ML systems

作者： Prashanth H C Yash Dharmesh Mogal Madhav Rao International Institute of Information Technology Bangalore IN International Institute of information Technology Bangalore IN

ISBN: (纸本)9798400716492

Systolic array (SA) architectures are suitable hardware accelerators to run artificial intelligence (AI) algorithms due to their scalable and pipelined structure. Inferencing on edge devices enables decision-making capability in a small, power-efficient form factor suitable for many IoT-based applications. Convolutional neural networks (CNNs), which form a major part of AI algorithms, are generally memory and compute-intensive algorithms. Optimizing resource-constrained SA-based edge-computing system architecture is crucial. The primary challenge is to find an optimal SA configuration with sufficient on-chip and off-chip memory, processing elements, and optimal memory-bandwidth interface for achieving minimum latency and stall cycles while maximizing compute utilization and mapping efficiency. This work characterizes the popular ResNet50 CNN model on different SA architecture configurations that are structured using a realistic designer tool SCALE-SIM. The design-space exploration of SA was carried out to run the Resnet50 model by configuring i) array sizes (processing elements matrix), ii) SRAM memory sizes available to the computing elements, iii) varying interface bandwidth of memory to the computing elements. The compute latency in clock cycles, SA utilization, Mapping efficiency, number of Stall cycles, and SRAM and DRAM bandwidth requirements were extracted using cycle accurate runs. All three types of data-flow models, including output stationary (OS), weight stationary (WS), and input stationary (IS), were investigated in this work. Unlike previous research that relied on benchmarking various mathematical models, this study focuses on characterizing SA configurations using the ResNet50 model. It presents comprehensive results for all the investigated configurations and establishes the most efficient SA architecture for running CNN architectures. Moreover, it provides valuable insights for designing resource-constrained edge-inferencing systems.

关键词： AI Accelerators CNN Design Space Exploration Systolic Array

来源：评论

学校读者我要写书评

暂无评论

Machine Learning Algorithm Based Plant Root Contour Extraction in Edge devices

Machine Learning Algorithm Based Plant Root Contour Extracti...

引用

Progress in Electromagnetic Research Symposium (PIERS)

作者： Peng Zhao Ting Fang Tan Lihui Wang School of Information and Communication Engineering Hainan University Haikou China

ISBN: (数字)9798350375909

ISBN: (纸本)9798350375916

The status of plant roots serves as a crucial metric reflecting the plant’s growth state. Extracting and processing the contour of plant roots aids in analyzing the growth status effectively. In this study, high-resolution root images of plants were obtained using embedded optical imaging devices, transferred to edge computing nodes for image processing. By fine-tuning machine learning parameters, an artificial neural network model was developed to integrate Holistically-Nested Edge Detection (HED) and Canny edge detection *** paper presents a methodology for model training and application based on convolutional neural networks on edge devices, typically single-board computers with limited memory (such as Raspberry Pi), addressing issues with traditional Canny edge detection’s sensitivity to salt-and-pepper noise and susceptibility to producing false contours due to image gradient changes. The algorithm leverages HED, employing machine learning solutions through a large collection of plant root image data for pre-training the model, effectively filtering images while retaining accurate edge information. Furthermore, the output of this method serves as input for the Canny edge extraction algorithm, utilizing an adaptive thresholding algorithm to generate clear plant root contour *** results demonstrate that the combined approach using embedded artificial neural network models and Canny contour extraction reduces device performance requirements while improving the quality of root contour extraction. This research contributes to a more accurate understanding of morphological features of plant root systems, providing comprehensive imaging data for plant root studies.

关键词： Training Adaptation models Machine learning algorithms Accuracy Thresholding (Imaging) Image edge detection Computational modeling

来源：评论

学校读者我要写书评

暂无评论

Adaptive exposure control for image-based visual-servo systems using local gradient information

引用

JOURNAL OF THE optical SOCIETY OF AMERICA A-OPTICS IMAGE SCIENCE AND VISION 2020年第1期37卷 56-62页

作者： Ding, Zhushun Chen, Xin Jiang, Zhe Tan, Cheng Natl Univ Def Technol Dept Aerospace & Engn Changsha 410073 Peoples R China Beijing Inst Aerosp Control Devices Beijing 100039 Peoples R China

We present anew method for automatic adjustment of camera exposure time for visual-servo systems. The proposed method can improve the robustness of image processing in a high-dynamic-range environment. In this paper, we evaluate an appropriate exposure time by computing the local gradient information of a target area, allowing a camera to capture images without losing target features under artificially adjusted illumination conditions. To validate the advantage of the proposed method, an evaluation is made using an off-the-shelf visual-servo system with a machine vision camera. The experimental results demonstrate the effectiveness of our method, which improves the performance of the target recognition algorithm. (C) 2019 optical Society of America

关键词： Image metrics Image processing Image processing algorithms Image quality Liquid crystals Machine vision

来源：评论

学校读者我要写书评

暂无评论

RESIST: Resilient Decentralized Learning Using Consensus Gradient Descent

arXiv

引用

arXiv 2025年

作者： Fang, Cheng Dixit, Rishabh Bajwa, Waheed U. Gürbüzbalaban, Mert Department of Electrical and Computer Engineering Rutgers University New BrunswickNJ United States Department of Mathematics University of California San DiegoCA United States Departments of Electrical and Computer Engineering and Statistics Rutgers University New BrunswickNJ United States Departments of Electrical and Computer Engineering Management Science and Information Systems and Statistics Rutgers University New BrunswickNJ United States

Empirical risk minimization (ERM) is a cornerstone of modern machine learning (ML), supported by advances in optimization theory that ensure efficient solutions with provable algorithmic convergence rates, which measure the speed at which optimization algorithms approach a solution, and statistical learning rates, which characterize how well the solution generalizes to unseen data. Privacy, memory, computational, and communications constraints increasingly necessitate data collection, processing, and storage across network-connected devices. In many applications, these networks operate in decentralized settings where a central server cannot be assumed, requiring decentralized ML algorithms that are both efficient and resilient. Decentralized learning, however, faces significant challenges, including an increased attack surface for adversarial interference during decentralized learning processes. This paper focuses on the man-in-the-middle (MITM) attack, wherein adversaries exploit communication vulnerabilities between devices to inject malicious updates during training, potentially causing models to deviate significantly from their intended ERM solutions. To address this challenge, we propose RESIST (Resilient dEcentralized learning using conSensus gradIent deScenT), an optimization algorithm designed to be robust against adversarially compromised communication links, where transmitted information may be arbitrarily altered before being received. Unlike existing adversarially robust decentralized learning methods, which often (i) guarantee convergence only to a neighborhood of the solution, (ii) lack guarantees of linear convergence for strongly convex problems, or (iii) fail to ensure statistical consistency as sample sizes grow, RESIST overcomes all three limitations. It achieves algorithmic and statistical convergence for strongly convex, Polyak–Lojasiewicz, and nonconvex ERM problems by employing a multistep consensus gradient descent framework and robust statis

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Tracking Quantum Coherence in Polariton Condensates with Time-Resolved Tomography

引用

Physical Review Letters 2023年第11期130卷 113601-113601页

作者： Carolin Lüders Matthias Pukrop Franziska Barkhausen Elena Rozas Christian Schneider Sven Höfling Jan Sperling Stefan Schumacher Marc Aßmann Experimentelle Physik 2 Technische Universität Dortmund D-44221 Dortmund Germany Department of Physics and Center for Optoelectronics and Photonics Paderborn (CeOPP) Universität Paderborn 33098 Paderborn Germany Institute of Physics University of Oldenburg D-26129 Oldenburg Germany Technische Physik Physikalisches Institut and Würzburg-Dresden Cluster of Excellence ct.qmat Universität Würzburg 97074 Würzburg Germany Theoretical Quantum Science Institute for Photonic Quantum Systems (PhoQS) Paderborn University Warburger Straße 100 33098 Paderborn Germany Wyant College of Optical Sciences University of Arizona Tucson Arizona 85721 USA

Long-term quantum coherence constitutes one of the main challenges when engineering quantum devices. However, easily accessible means to quantify complex decoherence mechanisms are not readily available, nor are sufficiently stable systems. We harness novel phase-space methods—expressed through non-Gaussian convolutions of highly singular Glauber-Sudarshan quasiprobabilities—to dynamically monitor quantum coherence in polariton condensates with significantly enhanced coherence times. Via intensity- and time-resolved reconstructions of such phase-space functions from homodyne detection data, we probe the systems’ resourcefulness for quantum information processing up to the nanosecond regime. Our experimental findings are confirmed through numerical simulations, for which we develop an approach that renders established algorithms compatible with our methodology. In contrast to commonly applied phase-space functions, our distributions can be directly sampled from measured data, including uncertainties, and yield a simple operational measure of quantum coherence via the distribution’s variance in phase. Therefore, we present a broadly applicable framework and a platform to explore time-dependent quantum phenomena and resources.

关键词： Exciton polariton Quantum information with hybrid systems Quantum interference effects Quantum optics Resource theories Semiconductor quantum optics Polariton condensate Semiconductor microcavities Homodyne & heterodyne detection Phase space dynamics Phase space methods

来源：评论

学校读者我要写书评

暂无评论

Quantum neural compressive sensing for ghost imaging

arXiv

引用

arXiv 2025年

作者： Zhai, Xinliang Xiao, Tailong Huang, Jingzheng Fan, Jianping Zeng, Guihua State Key Laboratory of Advanced Optical Communication Systems and Networks Institute for Quantum Sensing and Information Processing Shanghai Jiao Tong University Shanghai200240 China Hefei National Laboratory Hefei230088 China Shanghai Research Center for Quantum Sciences Shanghai201315 China AI Lab Lenovo Research Beijing100094 China

Demonstrating the utility of quantum algorithms is a long-standing challenge, where quantum machine learning becomes one of the most promising candidate that can be resorted to. In this study, we investigate a quantum neural compressive sensing algorithm for ghost imaging to showcase its utility. The algorithm utilizes the variational quantum circuits to reparameterize the inverse problem of ghost imaging and uses the inductive bias of the physical forward model to perform optimization. To validate the algorithm’s effectiveness, we conduct optical ghost imaging experiments, capturing signals from objects at different physical sampling rates and detection signal-to-noise ratios. The experimental results show that our proposed algorithm surpasses conventional methods in both visual appearance and quantitative metrics, achieving state-of-the-art performance. Importantly, we observe that the quantum neural network, guided by prior knowledge of physics, effectively overcomes the challenge of barren plateau in the optimization process. The proposed algorithm demonstrates robustness against various quantum noise levels, making it suitable for near-term quantum devices. Our study leverages physical inductive bias guided variational quantum algorithm, underscoring the potential of quantum computation in tackling a broad range of optimization and inverse problems. Copyright © 2025, The Authors. All rights reserved.

关键词： Inverse problems

来源：评论

学校读者我要写书评

暂无评论

先进引信技术的发展与展望

引用

科学通报 2023年第25期68卷 3293-3310页

作者：张合戴可人南京理工大学机械工程学院南京210094

引信是集通信交互、传感探测、信息处理、控制执行等诸多功能于一体的高密度集成小型光机电系统,是武器装备弹药的“大脑”,是决定弹药毁伤效果的关键.先进引信技术始终是世界主要国家大力发展和激烈竞争的关键技术领域.本文首先对先进... 详细信息

引信是集通信交互、传感探测、信息处理、控制执行等诸多功能于一体的高密度集成小型光机电系统,是武器装备弹药的“大脑”,是决定弹药毁伤效果的关键.先进引信技术始终是世界主要国家大力发展和激烈竞争的关键技术领域.本文首先对先进引信的技术研究进行了综述,介绍了近10年来引信在信息交联、末端探测、目标识别、瞬态控制四大基础方向和微小型化、多功能化、多源复合抗干扰三大应用方向的技术演进.在此基础上,本文提出了未来先进引信技术发展的五个战略方向:基于全寿命安全设计提高安全、基于抗干扰量子探测提高可靠性、基于快激活高能量电源提高动力、基于专用类脑架构处理器提高算力、基于引信战斗部一体化设计提高毁伤效能,为未来引信技术的发展提供参考借鉴.

关键词：先进引信信息交联瞬态控制多源复合全寿命安全

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：