检索结果-内蒙古大学图书馆

SSRN 2024年

作者： Xie, Hongbin Zhang, Haoran Song, Xuan Shi, Zhuoran Song, Ge Peng, Jinquan Xie, Yi Wang, Zhongyi Shenzhen China Shenzhen China Guangdong Provincial Key Laboratory of Brain-inspired Intelligent Computation Department of Computer Science and Engineering Southern University of Science and Technology Shenzhen China School of Urban Planning and Design Peking University No.2199 Lishui Road Nanshan District Guangdong Shenzhen518055 China Huawei Technologies CO. Ltd. Shenzhen China

In order to further reduce environmental pollution, many countries around the world are advocating and implementing the policies of "carbon peaking" and "carbon neutralization". Low-carbon Smart Mobility (LSM) is an innovative and environmentally-friendly transportation development model proposed over the last few years. By effectively integrating transportation resources and providing comfortable and intelligent transportation services, LSM can improve people's acceptance and utilization of electric transportation and public transportation, so as to reduce greenhouse gas emissions. At present, the development of LSM is still in its infancy, and most researches on LSM focus on the integration at the service level, without sufficient effective collection and analysis of various traffic data generated by massive traffic equipments at the bottom layer, so there is a lack of traffic value-added services based on data analysis. To solve the problem above, we proposes a novel construction scheme of building a data-driven Low-carbon Smart Mobility system. The scheme mainly realizes the interconnection of traffic equipment and the collection of traffic data in the whole transportation system by introducing Internet of Things (IoT) technology. On the basis of this, it further realizes the analysis and decision-making of massive traffic data by using cutting-edge science and technology such as big data, artificial intelligence, large language models (LLM), cloud computing and blockchain, so as to provide users with more environmentally-friendly, intelligent, convenient and comfortable transportation services, accelerate the development of urban green transportation, and ultimately achieve the goals of "carbon peak" and "carbon neutralization" as soon as possible. At the end of this article, we conducted a numerical experiment that demonstrates a case of scheduling decision in data-driven management for an automotive charging battery supply chain, thereby illustrating the feas

关键词： Internet of things

来源：评论

学校读者我要写书评

暂无评论

TransHD: Spatial Transformer Features Extraction for HDC Synergetic Learning

IEEE Transactions on Artificial Intelligence

引用

IEEE Transactions on Artificial Intelligence 2025年

作者： Hassan, Eman Bettayeb, Meriem Halawani, Yasmin Genssler, Paul R. Tesfai, Huruy Zweiri, Yahya Amrouch, Hussam Hadjileontiadis, Leontios J. Saleh, Hani Mohammad, Baker Khalifa University System on Chip Lab Department of Computer and Information Engineering Abu Dhabi 127788 United Arab Emirates Abu Dhabi University Department of Computer Science and Information Technology College of Engineering Abu Dhabi United Arab Emirates University of Dubai College of Engineering and IT Dubai United Arab Emirates Munich Institute of Robotics and Machine Intelligence Technical University of Munich TUM School of Computation Information and Technology Chair of AI Processor Design Munich Germany Khalifa University of Science and Technology Department of Aerospace Engineering Advanced Research and Innovation Center (ARIC) Abu Dhabi United Arab Emirates University of Stuttgart Semiconductor Test and Reliability (STAR) Stuttgart Germany Khalifa University of Science and Technology Department of Biomedical Engineering Abu Dhabi United Arab Emirates Aristotle University of Thessaloniki Department of Electrical and Computer Engineering Thessaloniki Greece

Artificial intelligence (AI) relies on pattern recognition and classification algorithms to achieve accurate and efficient decision-making. Convolutional neural networks (CNNs) are the state-of-the-art for processing 2D image data, but their high computational and data requirements limit their use in resource-constrained environments. Hyperdimensional computing (HDC) offers a lightweight alternative, excelling in 1D tasks, but struggles to achieve competitive accuracy for 2D image classification. Hybrid frameworks combining HDC with CNNs have been proposed to improve accuracy but inherit the computational demands of deep learning, making them unsuitable for edge and IoT devices. To address this, we propose TransHD, a framework integrating lightweight spatial transformer networks (STNs) with HDC to enhance image classification performance while maintaining efficiency. TransHD achieves up to 9% higher accuracy than base HDC models on MNIST and Fashion-MNIST using only 30% of the training data, and reduces computational complexity by 2.5x through optimized STN feature map usage. On resource-constrained platforms like the Raspberry Pi 4, TransHD accelerates inference times by 3.4x and improves energy efficiency by 3.2x, with an accuracy trade-off of approximately 3% compared to CNNs. This study demonstrates the potential of combining HDC with STNs to develop efficient AI solutions for IoT and edge computing, where low energy consumption and computational efficiency are critical. © 2020 IEEE.

关键词： Brain-inspired Computing Hyperdimensional Computing Image Classification Spatial Transformers

来源：评论

学校读者我要写书评

暂无评论

Immiscible-interface assisted direct metal drawing 29

Immiscible-interface assisted direct metal drawing

引用

29th Annual International Solid Freeform Fabrication Symposium - An Additive Manufacturing Conference, SFF 2018

作者： He, Li Fei, Fan Wang, Wenbo Song, Xuan Department of Industrial and Systems Engineering University of Iowa Iowa CityIA52242 United States Center for Computer Aided Design University of Iowa Iowa CityIA52242 United States

State-of-the-art metal additive manufacturing (AM) techniques construct a three-dimensional (3D) structure through sintering or melting dry metal powders in a layer-by-layer fashion, which consequently results in some typical manufacturing defects in the final structure, such as residual stress and highly-orientated microstructures. To overcome these defects, we present a new low-cost metal AM process, named Immiscible-interface assisted Direct Metal Drawing (II-DMD), which fabricates self-supported and isotropic 3D metal structures by continuously extruding a metal colloidal suspension within a second immiscible matrix colloidal suspension. The shape of the metal colloidal suspension is stabilized due to the presence of an immiscible interface between the two colloidal suspension systems. Dense metal structures can be achieved via post-consolidation of the self-stabilized metal-matrix systems, including liquid-phase drying and metal-phase sintering. In this article, the immiscible-interface-assisted self-stabilization mechanism is studied. The post-consolidation processes are discussed. Several test cases were fabricated and characterized. © Solid Freeform Fabrication 2018: Proceedings of the 29th Annual International Solid Freeform Fabrication Symposium - An Additive Manufacturing Conference, SFF 2018. All Rights Reserved.

关键词： Additives

来源：评论

学校读者我要写书评

暂无评论

Co-continuous Composite Material design Using the Volumetric Distance Field Based on Analytic Shape Functions

引用

Journal of Harbin Institute of Technology(New Series) 2019年第1期26卷 88-96页

作者： Huimin Han Dong-Jin Yoo Yan Liang Department of Computer Aided Mechanical Design Engineering Daejin University Department of Mechanical and Electrical Engineering Hainan Institute of Science and Technology

An effective and simple design method for co-continuous composite material construction is proposed by using a hybrid methodology with triply periodic minimal surface( TPMS) cellular topology and the volumetric distance field( VDF). After generating a set of VDF-based features for the given exterior shape and desired internal core structure,a series of simple modifications in distance fields enabled us to obtain an arbitrarily-shaped complex co-continuous composite material computational model. design results and manufactured prototypes through 3 D printing technology show that the proposed methodology has the potential to open a new paradigm for producing multifunctional next generation co-continuous composite materials which are impossible to design and manufacture using traditional CAD and CAM.

关键词： triply periodic minimal surface (TPMS) cellular topology volumetric distance field (VDF) co-continuous composite material model basic feature 3D printing technology

来源：评论

学校读者我要写书评

暂无评论

FDC: Feature Dropout Consistency for unsupervised domain adaptation semantic segmentation

引用

Neural Networks 2025年 189卷 107527页

作者： Rao, Chaoyu Fan, Wanshu Wang, Cong Yang, Xin Wei, Xiaopeng Zhou, Dongsheng National and Local Joint Engineering Laboratory of Computer Aided Design School of Software Engineering Dalian University China Department of Computing The Hong Kong Polytechnic University Hong Kong School of Computer Science and Technology Dalian University of Technology China

In Unsupervised Domain Adaptation Semantic Segmentation (UDASS), while self-training techniques have become one of the most effective methods to date, the absence of target labels makes models susceptible to overfitting. To address this problem, consistency techniques regularize the model by perturbing the target domain and enforcing consistent pixel predictions before and after the perturbation. Consistency techniques are based on the idea that a well-performing model should yield consistent predictions for the target data stream before and after perturbations. In this paper, we introduce the Feature Dropout Consistency (FDC) module, a novel approach that instigates perturbations at the feature level. This is achieved by implementing random feature dropout between the encoder and decoder of the student network. Subsequently, a feature dropout consistency loss is applied to minimize discrepancies in predictions between the perturbed student network and its teacher counterpart. Notably, FDC seamlessly integrates with existing self-training methodologies and, when paired with input perturbation, delves into an even broader perturbation spectrum. Through rigorous experimentation in standard UDA settings — training with synthetic labeled and real unlabeled data — FDC consistently outperforms baseline models, establishing new benchmarks on GTAV → Cityscapes and SYNTHIA → Cityscapes. © 2025

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

Analysis of Some Parameters of Synthesized Anodized Aluminum Oxide Matrices

Analysis of Some Parameters of Synthesized Anodized Aluminum...

引用

International Conference on Modern Problems of Radio engineering, Telecommunications and computer Science

作者： Roxoliana Bukliv Volodymyr Adamiv Ihor Bordun Nazariy Andrushchak Ihor Teslyuk Anatoliy Andrushchak Department of Chemistry and Technology of Inorganic Substances Institute of Chemistry and Chemical Technologies Lviv Polytechnic National University Lviv Ukraine O. G. Vlokh Institute of Physical Opics Ivan Franko National University of Lviv Lviv Ukraine Department of Electrical Power Engineering Czestochowa University of Technology Czestochowa Poland Department of Ecology and Sustainable Environmental Management Lviv Polytechnic National University Lviv Ukraine Computer-Aided Design Systems Lviv Politechnic National University Lviv Ukraine Department of Applied Physics and Nanomaterials Science Lviv Polytechnic National University Lviv Ukraine

ISBN: (数字)9798331520564

ISBN: (纸本)9798331520571

Studies have been conducted on anodized aluminum oxide (AAO) nanomatrices manufactured by SmartMembranes (SM) (Germany) and InRedox (USA). An X-ray fluorescence analyzer determined the quantitative and qualitative composition of AAO matrices. It is shown that the AAO matrices have a large transparency window (250 - 2750 nm) and the dependence of the second harmonic generation (SHG) signal on the pore size. The parameters of the porous structure of the samples were estimated by isothermal processes of nitrogen adsorption/desorption at its boiling point (T = 77 K).

关键词： Micrometers Sensitivity Fluorescence Market research Harmonic analysis Frequency conversion Telecommunications Nitrogen Isothermal processes Aluminum oxide

来源：评论

学校读者我要写书评

暂无评论

Temperature as a Chaotic Circuit Bifurcation Parameter

Temperature as a Chaotic Circuit Bifurcation Parameter

引用

2020 IEEE Conference of Russian Young Researchers in Electrical and Electronic engineering, EIConRus 2020

作者： Ostrovskii, Valerii Y. Nazare, Thalita E. Martins, Samir A. M. Nepomuceno, Erivelton G. Saint Petersburg Electrotechnical University LETI Department of Computer Aided Design Saint Petersburg Russia Federal University of São João Del-Rei Electrical Engineering Department São João del-Rei Brazil

ISBN: (纸本)9781728157610

The number of researches aimed at understanding the chaotic behavior of nonlinear dynamical systems has grown considerably in recent years. The development of electronic circuits that exhibit this type of behavior has been the interest of numerous works in the literature. Among the possible sets of nonlinear systems, the simplest in which one can observe bifurcation phenomena and chaotic behavior, followed by well- controlled experiments, are nonlinear electronic circuits. One of the most widely used tools for analysis and evaluation of chaotic behavior is known as the bifurcation diagram. Generally, in the analysis of these circuits, parameters such as voltage, current and frequency are used to verify their respective behaviors. Variable values of passive components such as resistors and capacitors are also widely used. The temperature has also been used as a bifurcation parameter in resistor, diode and inductor (RLD) circuits. However, there is little attention from the scientific community on temperature as a bifurcation parameter for electronic circuits using operational amplifiers such as the chaotic Jerk circuit. In this sense, this project aims to implement a chaotic Jerk circuit, composed of operational amplifiers, resistors and capacitors, and subject it to different temperature levels, using this variable as an analysis parameter. Thus, at the end of this work it was possible to verify that the temperature variation directly influences the behavior of the investigated system, thus reaching the final objective of the project, presenting that the temperature can be a bifurcation parameter for a chaotic Jerk circuit. © 2020 IEEE.

关键词： Resistors

来源：评论

学校读者我要写书评

暂无评论

Accelerating multi-objective neural architecture search by random-weight evaluation

arXiv

引用

arXiv 2021年

作者： Hu, Shengran Cheng, Ran He, Cheng Lu, Zhichao Wang, Jing Zhang, Miao Guangdong Provincial Key Laboratory of Brain-inspired Intelligent Computation Department of Computer Science and Engineering Southern University of Science and Technology Shenzhen518055 China Shanghai Aircraft Design and Research Institute Shanghai200135 China

For the goal of automated design of high-performance deep convolutional neural networks (CNNs), Neural Architecture Search (NAS) methodology is becoming increasingly important for both academia and industries. Due to the costly stochastic gradient descent (SGD) training of CNNs for performance evaluation, most existing NAS methods are computationally expensive for real-world deployments. To address this issue, we first introduce a new performance estimation metric, named Random-Weight Evaluation (RWE) to quantify the quality of CNNs in a cost-efficient manner. Instead of fully training the entire CNN, the RWE only trains its last layer and leaves the remainders with randomly initialized weights, which results in a single network evaluation in seconds. Second, a complexity metric is adopted for multi-objective NAS to balance the model size and performance. Overall, our proposed method obtains a set of efficient models with state-of-the-art performance in two real-world search spaces. Then the results obtained on the CIFAR-10 dataset are transferred to the ImageNet dataset to validate the practicality of the proposed algorithm. Moreover, ablation studies on NAS-Bench-301 datasets reveal the effectiveness of the proposed RWE in estimating the performance compared with existing methods. © 2021, CC BY.

关键词： Multiobjective optimization

来源：评论

学校读者我要写书评

暂无评论

Improving Chaotic Image Encryption Using Maps with Small Lyapunov Exponents

Improving Chaotic Image Encryption Using Maps with Small Lya...

引用

2020 Moscow Workshop on Electronic and Networking Technologies, MWENT 2020

作者： Santos, Thiago A. Magalhaes, Eduardo P. Basilio, Nayara P. Nepomuceno, Erivelton G. Karimov, Timur I. Butusov, Denis N. Federal University of São João Del-Rei Electrical Engineering Department São João del-Rei Brazil Petersburg Electrotechnical University Computer Aided Design Department Saint-Petersburg Russia Petersburg Electrotechnical University Youth Research Institute Saint-Petersburg Russia

ISBN: (纸本)9781728125718

Chaos-based encryption is one of the promising cryptography techniques that can be used. Although chaos-based encryption provides excellent security, the finite precision of number representation in computers affects decryption accuracy negatively. In this paper, a way to mitigate some problems regarding finite precision is analyzed. We show that the use of maps with small Lyapunov exponents can improve the performance of chaotic encryption scheme, making it suitable for image encryption. © 2020 IEEE.

关键词： Cryptography

来源：评论

学校读者我要写书评

暂无评论

Privacy-preserving Continual Federated Clustering via Adaptive Resonance Theory

arXiv

引用

arXiv 2023年

作者： Masuyama, Naoki Nojima, Yusuke Toda, Yuichiro Loo, Chu Kiong Ishibuchi, Hisao Kubota, Naoyuki The Graduate School of Informatics Department of Core Informatics Osaka Metropolitan University 1-1 Gakuen-cho Naka-ku Sakai-Shi Osaka599-8531 Japan The Faculty of Environmental Life Natural Science and Technology Okayama University 3-1-1 Tsushima-Naka Kita-ku Okayama Okayamashi700-8530 Japan The Department of Artificial Intelligence Faculty of Computer Science and Information Technology Universiti Malaya Kuala Lumpur50603 Malaysia The Guangdong Provincial Key Laboratory of Brain-inspired Intelligent Computation Department of Computer Science and Engineering Southern University of Science and Technology Shenzhen518055 China Graduate School of Systems Design Department of Mechanical Systems Engineering Tokyo Metropolitan University 6-6 Asahigaoka Hino-Shi Tokyo191-0065 Japan

With the increasing importance of data privacy protection, various privacy-preserving machine learning methods have been proposed. In the clustering domain, various algorithms with a federated learning framework (i.e., federated clustering) have been actively studied and showed high clustering performance while preserving data privacy. However, most of the base clusterers (i.e., clustering algorithms) used in existing federated clustering algorithms need to specify the number of clusters in advance. These algorithms, therefore, are unable to deal with data whose distributions are unknown or continually changing. To tackle this problem, this paper proposes a privacy-preserving continual federated clustering algorithm. In the proposed algorithm, an adaptive resonance theory-based clustering algorithm capable of continual learning is used as a base clusterer. Therefore, the proposed algorithm inherits the ability of continual learning. Experimental results with synthetic and real-world datasets show that the proposed algorithm has superior clustering performance to state-of-the-art federated clustering algorithms while realizing data privacy protection and continual learning ability. The source code is available at https://***/Masuyama-lab/FCAC. © 2023, CC BY-NC-ND.

关键词： Clustering algorithms

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：