检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Goldenstein, Nelson Sulam, Jeremias Romano, Yaniv Department of Electrical and Computer Engineering The Technion Israel Biomedical Engineering Department The Mathematical Institute for Data Science Johns Hopkins University United States Department of Electrical and Computer Engineering The Department of Computer Science The Technion Israel

Sparse auto-encoders are useful for extracting low-dimensional representations from high-dimensional data. However, their performance degrades sharply when the input noise at test time differs from the noise employed during training. This limitation hinders the applicability of auto-encoders in real-world scenarios where the level of noise in the input is unpredictable. In this paper, we formalize single hidden layer sparse auto-encoders as a transform learning problem. Leveraging the transform modeling interpretation, we propose an optimization problem that leads to a predictive model invariant to the noise level at test time. In other words, the same pre-trained model is able to generalize to different noise levels. The proposed optimization algorithm, derived from the square root lasso, is translated into a new, computationally efficient auto-encoding architecture. After proving that our new method is invariant to the noise level, we evaluate our approach by training networks using the proposed architecture for denoising tasks. Our experimental results demonstrate that the trained models yield a significant improvement in stability against varying types of noise compared to commonly used architectures. Copyright © 2024, The Authors. All rights reserved.

关键词： Optimization

来源：评论

学校读者我要写书评

暂无评论

A 145-ns Response Time High-Side Current Sense for a 100-V Monolithic GaN Half Bridge

A 145-ns Response Time High-Side Current Sense for a 100-V M...

引用

IEEE International Symposium on Circuits and Systems (ISCAS)

作者： Tommaso Francesco Tardani Andrea Gambero Sandro Rossi Antonio Aprile Edoardo Bonizzoni Piero Malcovati Department of Electrical Computer and Biomedical Engineering University of Pavia Italy STMicroelectronics Castelletto Italy

ISBN: (数字)9798350356830

ISBN: (纸本)9798350356847

This paper presents a monolithic GaN current sense circuit designed for the high-side of a 100-V half-bridge. The circuit is engineered to minimize current offsets and response times across a wide range of input voltages (10 V to 100 V), as well as varying temperatures and current loads. It is capable of tracking inductor current ripples and detecting the zero-current condition of the half-bridge high-side power transistor. The proposed solution demonstrates a simulated loop gain exceeding 100 dB and a bandwidth of 43.9 MHz, which achieves an accuracy of 99.4% and a response time of 145 ns under operating conditions of 50-V input voltage, 10-A current load, and 27°C.

关键词： Temperature sensors Resistors Operational amplifiers Temperature distribution Shunts (electrical) Accuracy Power transistors Sensors Time factors Inductors

来源：评论

学校读者我要写书评

暂无评论

Smart Vending Machines: Improving Service Quality with Integrated Controllers and SVM Algorithm 6

Smart Vending Machines: Improving Service Quality with Integ...

引用

6th International Conference on Energy, Power and Environment, ICEPE 2024

作者： Rajesh, K.S. Kalaiselvi, S. Sathyanathan, P. Sudha, K. Kavitha, G. Murugan, S. Madanapalle Institute of Technology & Science Department of Electrical and Electronics Engineering Andhra Pradesh Madanapalle India Kongu Engineering College Department of Computer Technology Tamil Nadu Perundurai India Vel Tech Rangarajan Dr.Sagunthala R&d Institute of Science and Technology Department of Electrical and Electronics Engineering Tamil Nadu Chennai India Department of Business Administration Tamil Nadu Mannargudi India Vinayaka Missions Kirupananda Variyar Engineering College Department of Computer Science and Engineering Tamil Nadu Salem India Saveetha University Saveetha School of Engineering Saveetha Institute of Medical and Technical Sciences Department of Biomedical Engineering Tamil Nadu Chennai India

ISBN: (纸本)9798350379297

The retail industry is transforming with the emergence of the Internet of Things (IoT) and embedded controllers with Support Vector Machines (SVM) algorithm. This research explores the advancements in vending machines, which combine technology and embedded controllers with traditional vending machines. These intelligent devices can collect data, analyze it, and make real-time decisions thanks to their sensors, touch screens, and sophisticated embedded controllers. Through this innovation, smart vending machines can offer consumers a shopping experience based on their purchases and interactions with the machine. With data analysis capabilities, these integrated controllers enable the machines to provide discounts, special offers, and even recommendations on what to buy. In addition, intelligent vending machines help streamline inventory management, preventing shortages of popular goods. Because of this, embedded controllers are transforming vending machine operations' future while radically altering the retail industry. By collecting and analyzing data in real-time, smart vending machines allow for data-driven decision-making, which improves inventory management and reduces waste. Since all payment methods are linked flawlessly, purchasers can be sure that their transactions will go through without a problem. © 2024 IEEE.

关键词： Inventory control

来源：评论

学校读者我要写书评

暂无评论

Multi-Context Recommendation Systems (CARS) in Autonomous Driving and Other Applications

Multi-Context Recommendation Systems (CARS) in Autonomous Dr...

引用

2022 International Telecommunications Conference, ITC-Egypt 2022

作者： Ali, Abdullahi Kashef, Rasha Ibrahim, Ahmed Fakhry Toronto Metropolitan University Electrical Computer and Biomedical Engineering Faculty of Engineering and Architectural Science TorontoON Canada Boston University Computer Science Department United States

ISBN: (数字)9781665488082

ISBN: (纸本)9781665488082

Recommendation systems (RS) are important for their instantaneous ability to suggest to users their desired items and for ensuring a smooth and worthwhile user experience. To optimize the user's experience with the RS, it can be programmed to be aware of the various contexts the user has while interacting with the system. Context-aware recommendation systems (CARS) aim to add more nuance to the process of predicting items by considering different factors such as location and time. This paper presents the foundations of recommendation systems, such as categories, evaluation metrics, datasets, and challenges. In this paper, we show the effectiveness of multi-contexts in autonomous driving. Three CARS models are developed and tested. Experimental results show that adding multi-contexts provides more personalized options for drivers, allowing intelligent decisions. © 2022 IEEE.

关键词： Recommender systems

来源：评论

学校读者我要写书评

暂无评论

SFA-Net: A Selective Features Absorption Network for Object Detection in Rainy Weather Conditions

引用

IEEE Transactions on Neural Networks and Learning Systems 2023年第8期34卷 5122-5132页

作者： Huang, Shih-Chia Hoang, Quoc-Viet Le, Trung-Hieu National Taipei University of Technology Department of Electronic Engineering Taipei10608 Taiwan National Taipei University of Technology International Graduate Program in Electrical Engineering and Computer Science Taipei10608 Taiwan Hung Yen University of Technology and Education Faculty of Information Technology Hungyen 17000 Viet Nam

In recent years, object detection approaches using deep convolutional neural networks (CNNs) have derived major advances in normal images. However, such success is hardly achieved with rainy images due to lack of visibility. Aiming to bridge this gap, in this article, we present a novel selective features absorption network (SFA-Net) to improve the performance of object detection not only in rainy weather conditions but also in favorable weather conditions. SFA-Net accomplishes this objective by utilizing three subnetworks, where the feature selection subnetwork is concatenated with the object detection subnetwork through the feature absorption subnetwork to form a unified model. To promote further advancement in object detection impaired by rain, we propose a large-scale rainy image dataset, named srRain, which contains both synthetic rainy images and real-world rainy images for training and testing purposes. srRain is comprised of 25 900 rainy images depicting diverse driving scenarios in the presence of rain with a total of 181 164 instances interpreting five common object categories. Experimental results display that our SFA-Net reaches the highest mean average precision (mAP) of 77.53% on a normal image set, 62.52% on a synthetic rainy image set, 37.34% on a collected natural rainy image set, and 32.86% on a published real rainy image set, surpassing current state-of-the-art object detectors and the combination of image deraining and object detection models while retaining a high speed. © 2012 IEEE.

关键词： Feature extraction

来源：评论

学校读者我要写书评

暂无评论

Robust Multiresolution and Multistain Background Segmentation in Whole Slide Images 23rd

Robust Multiresolution and Multistain Background Segmentatio...

引用

Proceedings of the 23rd Polish Conference on Biocybernetics and biomedical engineering, PCBEE 2023

作者： Jurgas, Artur Wodzinski, Marek Atzori, Manfredo Müller, Henning Information Systems Institute Sierre Switzerland Faculty of Electrical Engineering Automatics Computer Science and Biomedical Engineering AGH University of Science and Technology Krakow Poland Department of Neuroscience University of Padova Padova Italy Medical Faculty University of Geneva Geneva Switzerland

ISBN: (纸本)9783031384295

Background segmentation is an important step in analysis of histopathological images. It allows one to remove irrelevant regions and focus on the tissue of interest. However, background segmentation is challenging due to the variability of stain colors and intensity levels across different images, modalities, and magnification levels. In this paper, we present a learning-based model for histopathology background segmentation based on convolutional neural networks. We compare two multiresolution approaches to deal with the variability of magnification in histopathology images: (i) model that uses upscaling of smaller patches of the image, and (ii) model simultaneously trained on multiple resolution levels. Our model is characterized by solid performance both in multiresolution and multistain dyes (H &E and IHC), achieving good performance on publicly available dataset. The quantitative scores are, in terms of the Dice score, close to 94.71. The qualitative analysis presents strong performance on previously unseen cases from different distributions and various dyes. We freely release the model, weights, and ground-truth annotations to promote the open science and reproducible research. © 2024, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Pathology

来源：评论

学校读者我要写书评

暂无评论

Efficient storage of multidimensional telecom photons in a solid-state quantum memory

Optica Quantum

引用

Optica Quantum 2025年第3期3卷 295-302页

作者： Zongfeng Li Yisheng Lei Trevor Kling Mahdi Hosseini Department of Electrical and Computer Engineering and Applied Physics Program Northwestern University Evanston Illinois 60208 USA Elmore Family School of Electrical and Computer Engineering Purdue University West Lafayette Indiana 47907 USA

Efficient storage of telecom-band quantum optical information represents a crucial milestone for establishing distributed quantum optical networks. Erbium ions in crystalline hosts provide a promising platform for telecom quantum memories; however, their practical applications have been hindered by demanding operational conditions, such as ultra-high magnetic fields and ultra-low temperatures. In this work, we demonstrate the storage of telecom photonic qubits encoded in polarization, frequency, and time-bin bases. Using the atomic frequency comb protocol in an Er 3+ -doped crystal, we developed a memory initialization scheme that improves storage efficiency by over an order of magnitude. The observed results were made possible by the deliberate selection of the pumping sequence and the minimization of lattice interactions, to the extent possible without the use of dilution refrigerators or superconducting magnets.

关键词： Frequency combs Optical networks Photon polarization Quantum computation Quantum information Quantum memories

来源：评论

学校读者我要写书评

暂无评论

A Novel Strategy to Estimate and Manage the Power Consumption of Household Appliances 7

A Novel Strategy to Estimate and Manage the Power Consumptio...

引用

7th International Conference on Inventive Computation Technologies, ICICT 2024

作者： Mano Jemila, M.R. Pushpalatha, K.S. Mithuna, H.R. Supriya, C. Maheswari, K.T. Rajendiran, M. Dr. N. G.P. Institute of Technology Department of Biomedical Engineering Tamil Nadu Coimbatore641048 India Acharya Institute of Technology Department of Information Science And Engineering Karnataka Bengaluru560107 India Bannari Amman Institute of Technology Department of Electrical And Electronics Engineering Tamil Nadu Sathyamangalam638401 India Panimalar Engineering College Department of Computer Science And Engineering Tamil Nadu Chennai600123 India

ISBN: (纸本)9798350359299

Residential and commercial buildings account for approximately 26.6% and 21.2% of worldwide electrical power consumption, respectively. Because of the increasing need for energy and the lack of present energy sources, there will be a significant challenge in the future. As a result, with the rise of Smart Cities and Smart Grids, it is becoming increasingly vital to have effective strategies for controlling and reducing the consumption of electricity. One method of achieving consumption management is to predict the power usage of household equipment. This study focuses on power usage prediction using data from the UCI repository. In the preprocessing step Checking for null values, eliminating outliers, and modifying data are performed. Following that, the essential features are selected for dimension reduction using the Principal Component Analysis (PCA) technique. The Deep Sequential Prediction Model (DSPM) is the suggested deep learning method for predicting power consumption. Finally the significance of feature extraction and the effectiveness of the proposed model is evaluated using the R2 score and MSE. The proposed model is compared with the standard Artificial Neural Network (ANN) model. To examine the importance of the PCA method, feature-extracted data (from a correlation plot) and pre-processed data (without feature extraction) are used. The simulation shows that power consumption prediction utilizing the PCA feature extraction module along with the proposed DSPM model provides an excellent R2 score. © 2024 IEEE.

关键词： Electric power utilization

来源：评论

学校读者我要写书评

暂无评论

Fast non-iterative algorithm for 3D point-cloud holography

引用

Optics Express 2023年第22期31卷 36468-36485页

作者： Ersaro, Nathan Tessema Yalcin, Cem Murray, Liz Kabuli, Leyla Waller, Laura Muller, Rikky Graduate Program in Bioengineering University of California Berkeley and University of California San Francisco BerkeleyCA94720 United States Department of Electrical Engineering and Computer Sciences University of California Berkeley BerkeleyCA94720 United States

Recently developed iterative and deep learning-based approaches to computer-generated holography (CGH) have been shown to achieve high-quality photorealistic 3D images with spatial light modulators. However, such approaches remain overly cumbersome for patterning sparse collections of target points across a photoresponsive volume in applications including biological microscopy and material processing. Specifically, in addition to requiring heavy computation that cannot accommodate real-time operation in mobile or hardware-light settings, existing sampling-dependent 3D CGH methods preclude the ability to place target points with arbitrary precision, limiting accessible depths to a handful of planes. Accordingly, we present a non-iterative point cloud holography algorithm that employs fast deterministic calculations in order to efficiently allocate patches of SLM pixels to different target points in the 3D volume and spread the patterning of all points across multiple time frames. Compared to a matched-performance implementation of the iterative Gerchberg-Saxton algorithm, our algorithm’s relative computation speed advantage was found to increase with SLM pixel count, reaching >100,000x at 512 × 512 array format. © 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement.

关键词： Iterative methods

来源：评论

学校读者我要写书评

暂无评论

SVR-Optimized ANN Model for Predicting Earthquake Risk in electrical Substations Based on Disaster Datasets in the Aceh Region, Indonesia

引用

IEEE Access 2025年 13卷 105027-105043页

作者： Elvy Sahnur Nasution Yuwaldi Away Syahrial Ira Devi Sara Andri Novandri Doctoral Program School of Engineering Universitas Syiah Kuala Banda Aceh Indonesia Department of Electrical Engineering Universitas Muhammadiyah Sumatera Utara Medan Indonesia Department of Electrical and Computer Engineering Universitas Syiah Kuala Banda Aceh Indonesia Department of Computer Engineering Universitas Serambi Mekkah Banda Aceh Indonesia

This paper proposes a predictive model based on an Artificial Neural Network (ANN) optimized with Support Vector Regression (SVR) to predict seismic parameters as part of earthquake risk mitigation for substations. The proposed model aims to improve the accuracy of seismic parameter predictions, including seismic energy, Peak Ground Acceleration (PGA), and Modified Mercalli Intensity (MMI). A case study was conducted in the Aceh region, Indonesia, which is located near the subduction zone between the Indo-Australian Plate and the Eurasian Plate, making it highly seismically active. Therefore, this model is developed to identify substations most at risk of earthquake impacts while supporting disaster mitigation planning through early warning systems. The dataset used in the modeling includes information on the epicenter and substation coordinates, earthquake magnitude, depth, and the distance between the two points. The ANN training process uses backpropagation with sigmoid and tanh activation functions. The ANN outputs are then used as input features for the SVR model, applying a Radial Basis Function (RBF) kernel to capture the non-linear relationships between variables. Evaluation results show that the hybrid SVR-ANN model significantly improves prediction accuracy compared to conventional ANN. This improvement is indicated by a reduction in error and an increase in the coefficient of determination ( $R^{2}$ ). The findings of this paper can serve as a foundation for developing more effective mitigation systems to reduce earthquake impacts on electrical infrastructure.

关键词： Earthquakes Substations Predictive models Computational modeling Artificial neural networks Electronics packaging Accuracy Data models Prevention and mitigation Earth

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：