检索结果-内蒙古大学图书馆

arXiv 2022年

作者： Kan, Shichao Liang, Yixiong Li, Min Cen, Yigang Wang, Jianxin He, Zhihai School of Computer Science and Engineering Central South University Hunan Changsha410083 China Institute of Information Science School of Computer and Information Technology Beijing Jiaotong University Beijing100044 China Beijing Key Laboratory of Advanced Information Science and Network Technology Beijing100044 China Department of Electrical and Electronic Engineering Southern University of Science and Technology Shenzhen China Pengcheng Lab Shenzhen518066 China

A fundamental challenge in deep metric learning is the generalization capability of the feature embedding network model since the embedding network learned on training classes need to be evaluated on new test classes. To address this challenge, in this paper, we introduce a new method called coded residual transform (CRT) for deep metric learning to significantly improve its generalization capability. Specifically, we learn a set of diversified prototype features, project the feature map onto each prototype, and then encode its features using their projection residuals weighted by their correlation coefficients with each prototype. The proposed CRT method has the following two unique characteristics. First, it represents and encodes the feature map from a set of complimentary perspectives based on projections onto diversified prototypes. Second, unlike existing transformer-based feature representation approaches which encode the original values of features based on global correlation analysis, the proposed coded residual transform encodes the relative differences between the original features and their projected prototypes. Embedding space density and spectral decay analysis show that this multi-perspective projection onto diversified prototypes and coded residual representation are able to achieve significantly improved generalization capability in metric learning. Finally, to further enhance the generalization performance, we propose to enforce the consistency on their feature similarity matrices between coded residual transforms with different sizes of projection prototypes and embedding dimensions. Our extensive experimental results and ablation studies demonstrate that the proposed CRT method outperform the state-of-the-art deep metric learning methods by large margins and improving upon the current best method by up to 4.28% on the CUB dataset. Copyright © 2022, The Authors. All rights reserved.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

A Clinical Decision Support System (CDSS) for Unbiased Prediction of Caesarean Section Based on Features Extraction and Optimized Classification

引用

Computational Intelligence and Neuroscience 2022年第0期2022卷 1901735页

作者： Javeed, Ashir Ali, Liaqat Mohammed Seid, Abegaz Ali, Arif Khan, Dilpazir Imrana, Yakubu Aging Research Center Karolinska Institute Solna Sweden Department of Electrical Engineering University of Science and Technology Bannu Bannu Pakistan Information & Computing Technology Division College of Science and Engineering Hamad Bin Khalifa University Doha Qatar Department of Computer Science University of Science and Technology Bannu Bannu Pakistan School of Engineering University for Development Studies Tamale Ghana Chengdu China

Nowadays, caesarean section (CS) is given preference over vaginal birth and this trend is rapidly rising around the globe, although CS has serious complications such as pregnancy scar, scar dehiscence, and morbidly adherent placenta. Thus, CS should only be performed when it is absolutely necessary for mother and fetus. To avoid unnecessary CS, researchers have developed different machine-learning- (ML-) based clinical decision support systems (CDSS) for CS prediction using electronic health record of the pregnant women. However, previously proposed methods suffer from the problems of poor accuracy and biasedness in ML. To overcome these problems, we have designed a novel CDSS where random oversampling example (ROSE) technique has been used to eliminate the problem of minority classes in the dataset. Furthermore, principal component analysis has been employed for feature extraction from the dataset while, for classification purpose, random forest (RF) model is deployed. We have fine-tuned the hyperparameter of RF using a grid search algorithm for optimal classification performance. Thus, the newly proposed system is named ROSE-PCA-RF and it is trained and tested using an online CS dataset available on the UCI repository. In the first experiment, conventional RF model is trained and tested on the dataset while in the second experiment, the proposed model is tested. The proposed ROSE-PCA-RF model improved the performance of traditional RF by 4.5% with reduced time complexity, while only using two extracted features through the PCA. Moreover, the proposed model has obtained 96.29% accuracy on training data while improving the accuracy of 97.12% on testing data. © 2022 Ashir Javeed et al.

关键词： Classification (of information)

来源：评论

学校读者我要写书评

暂无评论

Conducting Ethical Research on Emerging Technologies for Children

Conducting Ethical Research on Emerging Technologies for Chi...

引用

2025 CHI Conference on Human Factors in Computing Systems, CHI EA 2025

作者： Hourcade, Juan Pablo Bakala, Ewelina Bonsignore, Elizabeth Currin, Flannery Hope Fails, Jerry Alan Gilhoi, Amy Medina, Nuria Medina Norris, Delaney Onions, Meredith Pires, Ana Cristina Walsh, Greg Yarosh, Svetlana Yip, Jason University of Iowa Iowa CityIA United States INCO Facultad de Ingeniería UDELAR Montevideo Montevideo Uruguay College of Information Studies University of Maryland College ParkMD United States Denison University GranvilleOH United States Boise State University BoiseID United States Department of Computer Science The University of Iowa Iowa CityIA United States Universidad de Granada Granada Spain The University of Iowa Iowa CityIA United States The University of Iowa Iowa CityIA United States Interactive Technology Institute ITI LARSyS Instituto Superior Técnico Universidade de Lisboa Portugal Digital Whimsy Lab. University of Baltimore BaltimoreMD United States Department of Computer Science and Engineering University of Minnesota MinneapolisMN United States Information School University of Washington SeattleWA United States

ISBN: (纸本)9798400713958

This SIG will provide child-computer interaction researchers and practitioners, as well as other interested CHI attendees, an opportunity to discuss topics related to conducting ethical research on emerging technologies for children. While the community has extensively debated on ethical issues, we have not had ample discussion of how to ethically manage research on emerging technologies, often designed for adults, that may also be used by children or affect children. More specifically, we would like to discuss ethical aspects related to motivations for research, how research is conducted, and how it is reported. We hope these discussions will go well beyond legal requirements from ethics boards and help provide foundational guidance for research on emerging technologies with children, but also inform similar research with other vulnerable communities. © 2025 Copyright held by the owner/author(s).

关键词： Ethical technology

来源：评论

学校读者我要写书评

暂无评论

Design and Development of IoT Oriented Solar Powered Smart Home Controlling Mechanism

Design and Development of IoT Oriented Solar Powered Smart H...

引用

Sustainable Communication Networks and Application (ICSCNA), International Conference on

作者： R. Tamezheneal K. Kajendran J.C. Vinitha Hemalatha S K.B. Aruna V Samuthira Pandi Department of Electronics and Communication Engineering Sri Sai Ram engineering college Chennai Tamil Nadu India Department of Computer Science and Engineering Panimalar Engineering College Chennai Tamil Nadu India Department of Electrical and Electronics Engineering Rajalakshmi Engineering College Chennai Tamil Nadu India Department of Electrical and Electronics Engineering Prince Shri Venkateshwara Padmavathy Engineering College Chennai Tamil Nadu India Department of Computer Science and Engineering S.A.Engineering college Chennai Tamil Nadu India Centre for Advanced Wireless Integrated Technology Chennai Institute of Technology Chennai Tamil Nadu India

This research study addresses the need for renewable energy in residence by discussing the development and building of a solar assisted automation system for homes equipped with a wireless Internet of Things (IoT) enabled technology. The photovoltaic array powers the intelligent home using a highly efficient Maximum Power Point Tracking (MPPT) control system. The research's design, analysis, and methodology are all presented here, along with the study's findings on the market. The primary building blocks of a solar UPS are a solar panel and a custom-designed inverter circuit. The inverter circuit was designed in accordance with solar panel specs and requirements. Many other circuits served as inspiration for the present circuit's redesign. The development experts of currently available backup power supplies have also been consulted in order to plan for prospective replacements should a component fail or become unavailable. In this post, we'll look at some of the several ways a solar-powered UPS may be built and how its features might work. This suggests that solar UPSs may soon be able to compete with and even outperform their electrical counterparts. The key components of the design are an outside solar panel equipped with solar cells to convert solar energy towards electrical power and a power conversion circuit to convert the DC current into AC current for usage in household appliances. Two Energy throw electrical parts and one Arduino UNO controller with sensors were used in the development of this prototype automation system. In this paper, we propose developing and implementing a residence energy control system to help individuals and families keep tabs on their energy consumption and adjust their habits accordingly. The purpose of this project is to design, develop, and implement a residential energy optimization system that enables sensing, control, and intelligent algorithms to help clients learn more about and control their consumption of energy.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Modular reconfiguration of flexible production systems using machine learning and performance estimates

引用

IFAC-PapersOnLine 2022年第10期55卷 353-358页

作者： D. Scrimieri O. Adalat S. Afazov S. Ratchev Department of Computer Science University of Bradford Bradford BD7 1DP UK Department of Engineering Nottingham Trent University Nottingham NG11 8NS UK Institute for Advanced Manufacturing University of Nottingham Nottingham NG8 1BB UK

This paper presents an agent-based framework for reconfiguring modular assembly systems using machine learning and system performance estimates based on previous reconfigurations. During a reconfiguration, system integrators and engineers make changes to the machine to meet new production requirements by increasing capacity or manufacturing new product variants. The framework provides a method for automatically evaluating these changes in terms of impact on the performance of the production system, and building a knowledge base. Such knowledge is used to support future reconfigurations by recommending changes that are likely to improve the performance based on previous reconfigurations. The agent architecture of the framework has two levels, one for individual assembly stations and one for the entire production line. Knowledge bases of changes are built and utilised at both levels using machine learning and performance estimates. A prototype implementation of the proposed framework has been evaluated on an assembly production system in an industrial scenario. Preliminary results show that framework helps to reduce the time and resources required to complete a system reconfiguration and reach the desired production objectives.

关键词： Reconfiguration production systems assembly agents machine learning

来源：评论

学校读者我要写书评

暂无评论

Wander 22

Wander

引用

Proceedings of the SIGGRAPH Asia 2022 Art Gallery

作者： Yuqian Sun Chang Hee Lee Chenhang Cheng Ali Asadipour Ying Xu Yihua Li Computer Science Research Centre Royal College of Art London United Kingdom Affective Systems and Cognition Lab Industrial Design Department College of Engineering Korea Advanced Institute of Science and Technology Daejeon Republic of Korea Unaffiliated Shanghai China Department of Industrial Design School of Art and Design Wuhan University of Technology Wuhan China Department of Product Design College of Fashion and Design Donghua University Shanghai China

This is a public art project centred on an AI chatbot called Wander. She can generate interactive fiction based on realworld locations' knowledge graphs. Through public participation, the stories are visualised on...

ISBN: (纸本)9781450394710

关键词：

来源：评论

学校读者我要写书评

暂无评论

VirtualPose: Learning Generalizable 3D Human Pose Models from Virtual Data

arXiv

引用

arXiv 2022年

作者： Su, Jiajun Wang, Chunyu Ma, Xiaoxuan Zeng, Wenjun Wang, Yizhou Center for Data Science Peking University China Center on Frontiers of Computing Studies Peking University China Dept. of Computer Science Peking University China Inst. for Artificial Intelligence Peking University China Microsoft Research Asia China EIT Institute for Advanced Study United States

While monocular 3D pose estimation seems to have achieved very accurate results on the public datasets, their generalization ability is largely overlooked. In this work, we perform a systematic evaluation of the existing methods and find that they get notably larger errors when tested on different cameras, human poses and appearance. To address the problem, we introduce VirtualPose, a two-stage learning framework to exploit the hidden "free lunch" specific to this task, i.e. generating infinite number of poses and cameras for training models at no cost. To that end, the first stage transforms images to abstract geometry representations (AGR), and then the second maps them to 3D poses. It addresses the generalization issue from two aspects: (1) the first stage can be trained on diverse 2D datasets to reduce the risk of over-fitting to limited appearance;(2) the second stage can be trained on diverse AGR synthesized from a large number of virtual cameras and poses. It outperforms the SOTA methods without using any paired images and 3D poses from the benchmarks, which paves the way for practical applications. Code is available at https://***/wkom/VirtualPose. Copyright © 2022, The Authors. All rights reserved.

关键词： Cameras

来源：评论

学校读者我要写书评

暂无评论

Learning Markov models of fading channels in wireless control networks: A regression trees based approach

arXiv

引用

arXiv 2021年

作者： Reyes, Luis Felipe Florenzan Smarra, Francesco Lun, Yuriy Zacchia D’Innocenzo, Alessandro Department of Information Engineering Computer Science and Mathematics University of L’Aquila AQ67100 Italy IMT School for Advanced Studies Lucca LU55100 Italy

Finite-state Markov models are widely used for modeling wireless channels affected by a variety of nonidealities, ranging from shadowing to interference. In an industrial environment, the derivation of a Markov model based on the wireless communication physics can be prohibitive as it requires a complete knowledge of both the communication dynamics parameters and of the disturbances/interferers. In this work, a novel methodology is proposed to learn a Markov model of a fading channel via historical data of the signal-to-interference-plus-noise-ratio (SINR). Such methodology can be used to derive a Markov jump model of a wireless control network, and thus to design a stochastic optimal controller that takes into account the interdependence between the plant and the wireless channel dynamics. The proposed method is validated by comparing its prediction accuracy and control performance with those of a stationary finite-state Markov chain derived assuming perfect knowledge of the physical channel model and parameters of a WirelessHART point-to-point communication based on the IEEE-802.15.4 standard. Copyright © 2021, The Authors. All rights reserved.

关键词： Markov processes

来源：评论

学校读者我要写书评

暂无评论

Advantages and limitations of quantum routing

arXiv

引用

arXiv 2022年

作者： Bapat, Aniruddha Childs, Andrew M. Gorshkov, Alexey V. Schoute, Eddie Lawrence Berkeley National Laboratory BerkeleyCA94720 United States Joint Center for Quantum Information and Computer Science NIST University of Maryland College ParkMD20742 United States Joint Quantum Institute NIST University of Maryland College ParkMD20742 United States Department of Physics University of Maryland College ParkMD20742 United States Institute for Advanced Computer Studies University of Maryland College ParkMD20742 United States Department of Computer Science University of Maryland College ParkMD20742 United States Computer Computational and Statistical Sciences Division Los Alamos National Laboratory Los AlamosNM87545 United States

The Swap gate is a ubiquitous tool for moving information on quantum hardware, yet it can be considered a classical operation because it does not entangle product states. Genuinely quantum operations could outperform swap for the task of permuting qubits within an architecture, which we call routing. We consider quantum routing in two models: (i) allowing arbitrary two-qubit unitaries, or (ii) allowing Hamiltonians with norm-bounded interactions. We lower bound the circuit depth or time of quantum routing in terms of spectral properties of graphs representing the architecture interaction constraints, and give a generalized upper bound for all simple connected n-vertex graphs. In particular, we give conditions for a superpolynomial classical-quantum routing separation, which exclude graphs with a small spectral gap and graphs of bounded degree. Finally, we provide examples of a quadratic separation between gate-based and Hamiltonian routing models with a constant number of local ancillas per qubit and of an Ω(n) speedup if we also allow fast local interactions. Copyright © 2022, The Authors. All rights reserved.

关键词： Qubits

来源：评论

学校读者我要写书评

暂无评论

Comparison of Feature Extraction and Matching Techniques Applied to Monocular UAV Images With Low-quality and Land Cover Variations

Comparison of Feature Extraction and Matching Techniques App...

引用

IEEE Latin American Robotics Symposium, LARS

作者： Nathan A. Z. Xavier Jonathan De A. Lapa Douglas D. De C. Honório Rafael C. De Oliveira Elcio H. Shiguemori Marcos R. O. A. Maximo Marco A. P. Domiciano Technical College of UFMG Federal University of Minas Gerais (UFMG) Belo Horizonte MG Brazil Aeronautics Institute of Technology (ITA) São José dos Campos SP Brazil C4ISR Institute for Advanced Studies (IEAv) São José dos Campos SP Brazil Computer Science Division Aeronautics Institute of Technology (ITA) São José dos Campos SP Brazil

Extracting features from images is a recurring task in applications highly dependent on cameras, mainly on UAV visual navigation. By using sequential images, on different land covers and low-quality images, variable performance is expected based on both modification characteristics. This study applies the techniques GFTT, SIFT, MSER, FAST, BRISK, ORB, AKAZE, and SuperGlue on the dataset KDD-BR 2022, which contains multiple UAV images of various land covers and low-quality images. The performance of each technique on feature extracting and matching is presented while verifying the methods that better generalize this process.

关键词：

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：