检索结果-内蒙古大学图书馆

SFedXL: Semi-Synchronous Federated Learning With Cross-Sharpness and Layer-Freezing

IEEE Internet of Things Journal 2025年

作者： Zhao, Mingxiong Zhao, Shihao Feng, Chenyuan Yang, Howard H. Niyato, Dusit Quek, Tony Q. S. Yunnan University National Pilot School of Software Kunming650500 China Ministry of Education Engineering Research Center of Integration and Application of Digital Learning Technology Beijing100039 China Ministry of Education Engineering Research Center of Cyberspace Kunming650504 China Yunnan University of Finance and Economics Yunnan Key Laboratory of Service Computing Kunming650221 China EURECOM Sophia Antipolis06410 France Zhejiang University University of Illinois Urbana-Champaign Institute Zhejiang University Haining314400 China Nanyang Technological University School of Computer Science and Engineering 639798 Singapore Singapore University of Technology and Design Information Systems Technology and Design Pillar 487372 Singapore

Federated learning (FL) emerges as a potential solution for enabling multiple terminal devices to collaboratively accomplish computational tasks within an Unmanned Aerial Vehicle (UAV) swarm. However, traditional FL approaches, predicated on synchronous data aggregation, are not feasible for a UAV swarm owing to the inherently variable and dynamic nature of their communication networks compared with terrestrial systems. Furthermore, the data procured by UAVs is often highly heterogeneous, attributable to disparities in deployment environments and device attributes. Considering the distinct flight paths and unique operational conditions encountered by different UAVs, a considerable amount of data remains unlabeled. To tackle the challenges associated with asynchronous operations and the prevalence of unlabeled data, we introduce a novel framework termed Semi-synchronous FL with Cross-Sharpness and Layer-Freezing (SFedXL), tailored for a UAV swarm. In particular, we devise a cross-sharpness model training strategy aimed at optimizing the utilization of both labeled and unlabeled datasets. Additionally, we propose an innovative semi-synchronous model aggregation protocol, complemented by client-specific layer-freezing and client cluster scheduling, designed to expedite the training process. Our simulation results indicate that the proposed algorithm surpasses current FL methods in terms of object recognition accuracy and communication efficiency, albeit with a trade-off of increased local computation latency. © 2014 IEEE.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Efficient alltoall_inplace communication algorithms with limited buffer space

Efficient alltoall_inplace communication algorithms with lim...

引用

2022 International Conference on computer Application and Information Security, ICCAIS 2022

作者： Fang, Yanfei Li, Yanbing Dong, Enming Qi, Fengbin He, Wangquan Wang, Yunfei Xu, Zhengmao National Research Center of Parallel Computer Engineering and Technology Beijing100190 China

ISBN: (纸本)9781510663459

Alltoall is an important collective communication mode that is widely used in scientific and engineering computing applications. In applications such as FFT, using as much memory space as possible can improve applications' efficiency and expand the scale of problems to be solved. High-bandwidth alltoall_inplace communication under space-constrained buffer conditions can significantly improve the performance of such applications. In this paper, two alltoall_inpalce communication algorithms under the condition of limited buffer space are designed for the domestic parallel computing system based on Sunway many-core processor. The alltoall_inpalce communication interfaces for both the core array and the whole CPU node are realized. The experimental results show that algorithms implemented in this paper can achieve high bandwidth using limited buffer space. © 2023 SPIE.

关键词： Bandwidth

来源：评论

学校读者我要写书评

暂无评论

Semantic-Aware Prompt Learning for Multimodal Sarcasm Detection

Semantic-Aware Prompt Learning for Multimodal Sarcasm Detect...

引用

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Guangjin Wang Bao Wang Fuyong Xu Zhenfang Zhu Peipei Wang Ru Wang Peiyu Liu School of Information Science and Engineering Shandong Normal University Jinan China School of Information Science and Electrical Engineering Shandong Jiao Tong University Jinan China Key Laboratory of Computing Power Network and Information Security Ministry of Education Shandong Computer Science Center (National Supercomputer Center in Jinan) Qilu University of Technology (Shandong Academy of Sciences) Jinan China Shandong Provincial Key Laboratory of Industrial Network and Information System Security Shandong Fundamental Research Center for Computer Science Jinan China

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Multimodal sarcasm detection aims to identify whether utterances express sarcastic intentions contrary to their literal meaning based on multimodal information. However, existing methods fail to explore the model’s "ability to understand" the semantics expressed by sentences in the image context from semantic diversity perspectives. In this paper, we propose a multi-view semantic awareness method, which concretizes semantics from multiple perspectives to improve the model’s ability to capture different semantic features. Specifically, two learnable prefixes are attached to the text representation respectively to construct semantic representations from both the literal meaning and sarcastic intention perspectives. Then, image-text information is further fused through cross-attention to guide the semantic representation of different perspectives in the image context. Finally, the semantics expressed by prefixes are strengthened through KL divergence, thereby encouraging the model to capture two distinctive semantic features. Experiments on benchmark datasets demonstrate the effectiveness of our method.

关键词： Semantics Benchmark testing Speech enhancement Signal processing Vectors Acoustics Context modeling

来源：评论

学校读者我要写书评

暂无评论

Fostering artificial societies using social learning and social control in parallel emergency management systems

引用

Frontiers of computer Science 2012年第5期6卷 604-610页

作者： Wei DUAN Xiaogang QIU Research Center of Military Computational Experiments and Parallel Systems Technology College of Mechatronics Engineering and AutomationNational University of Defense TechnologyChangsha 410073China

How can we foster and grow artificial societies so as to cause social properties to emerge that are logical, consistent with real societies, and are expected by design- ers？ We propose a framework for fostering artificial soci- eties using social learning mechanisms and social control ap- proaches. We present the application of fostering artificial so- cieties in parallel emergency management systems. Then we discuss social learning mechanisms in artificial societies, in- cluding observational learning, reinforcement learning, imi- tation learning, and advice-based learning. Furthermore, we discuss social control approaches, including social norms, social policies, social reputations, social commitments, and sanctions.

关键词： artificial societies social computing social learning social control agent-based simulation

来源：评论

学校读者我要写书评

暂无评论

On-chip multi-timescale spatiotemporal optical synchronization

arXiv

引用

arXiv 2025年

作者： Xu, Lida Mehrabad, Mahmoud Jalali Flower, Christopher J. Moille, Gregory Restelli, Alessandro Suarez-Forero, Daniel G. Chembo, Yanne Mittal, Sunil Srinivasan, Kartik Hafezi, Mohammad Joint Quantum Institute Quantum Technology Center University of Maryland College ParkMD20742 United States Joint Quantum Institute University of Maryland National Institute of Standards and Technology College ParkMD20742 United States Department of Electrical and Computer Engineering Institute for Research in Electronics and Applied Physics Joint Quantum Institute University of Maryland College ParkMD20742 United States Department of Electrical and Computer Engineering Northeastern University BostonMA United States

Mode-locking mechanisms are key resources in nonlinear optical phenomena, such as micro-ring solitonic states, and have transformed metrology, precision spectroscopy, and optical communication. However, despite significant efforts, mode-locking has not been demonstrated in the independently tunable multi-timescale regime. Here, we vastly expand the nonlinear mode-locking toolbox into multi-timescale synchronization on a chip. We use topological photonics to engineer a 2D lattice of hundreds of coupled silicon nitride ring resonators capable of hosting nested mode-locked states with a fast (≈ 1 THz) single-ring and a slow (≈ 3 GHz) topological super-ring timescales. We demonstrate signatures of multi-timescale mode-locking including quadratic distribution of the pump noise with the two-time azimuthal mode dimensions, as expected by mode-locking theory. Our observations are further corroborated by direct signatures of the near-transform-limit repetition beats and the formation of the temporal pattern on the slow timescale. Moreover, we show that these exotic properties of edge-confined mode-locked states are in sharp contrast to bulk and single-ring counterparts and establish a clear pathway for their identification. Our unprecedented demonstration of mode-locking in topological combs unlocks the implementation of lattice-scale synchronization and independently tunable multi-timescale mode-locking phenomena, also the exploration of the fundamental nonlinearity-topology interplay on a chip. © 2025, CC BY.

关键词： Silicon nitride

来源：评论

学校读者我要写书评

暂无评论

Sunway supercomputer architecture towards exascale computing: analysis and practice

引用

Science China(Information Sciences) 2021年第4期64卷 177-197页

作者： Jiangang GAO Fang ZHENG Fengbin QI Yajun DING Hongliang LI Hongsheng LU Wangquan HE Hongmei WEI Lifeng JIN Xin LIU Daoyong GONG Fei WANG Yan ZHENG Honghui SUN Zhou ZHOU Yong LIU Hongtao YOU National Research Center of Parallel Computer Engineering and Technology

In recent years, the improvements of system performance and energy efficiency for supercomputers have faced increasing challenges, which create more intensive demands on the architecture design for realizing exascale computing. This paper first analyzes the main requirements of exascale computing on the aspects of the parallel computing application and supercomputing center operation. Afterwards, a mapping scheme of "demands-challenges-architecture" is proposed. Then, the major challenges of exascale supercomputer, such as scalability, power consumption, data movement, programming and availability, are thoroughly analyzed, and the corresponding appropriate solutions are proposed. Moreover, this paper proposes the Sunway computer architecture towards exascale computing in which the many-core processor, network chipset and software system are all domestically-designed. The technology roadmap of Sunway supercomputer will hold the comprehensive design methods for the architecture, including the processor, interconnect network,assembly structure, power supply, cooling system, system software, parallel algorithm and application support, promising great advances for exascale supercomputing.

关键词： supercomputer exascale Sunway scalability power consumption data movement programming availability

来源：评论

学校读者我要写书评

暂无评论

A survey on scalable consensus algorithms for blockchain technology

Cyber Security and Applications

引用

Cyber Security and Applications 2025年 3卷

作者： Jain, Ankit Kumar Gupta, Nishant Gupta, Brij B. National Institute of Technology Kurukshetra India Department of Computer Science and Information Engineering Asia University Taiwan Taichung 413 China Symbiosis Centre for Information Technology (SCIT) Symbiosis International University Pune India Center for Interdisciplinary Research University of Petroleum and Energy Studies (UPES) Dehradun India

The process of reaching an agreement on a value within a distributed network, known as a consensus problem, is a defining feature of blockchain. This consensus problem can be seen in various applications like load balancing, transaction validation in blockchain, and distributed computing. In recent years, many researchers have provided solutions to this problem. Hence we have presented a survey in which we delved into blockchain consensus algorithms and conducted a comparative analysis of all the consensus algorithms to provide information about each protocol's advantages and drawbacks. This survey starts with the standard proof-of-work consensus protocol applied in bitcoin cryptocurrency and its limitations on the ground of the following parameters: throughput (transactions per second), latency, forks, fault tolerance, double spending attacks, and power consumption. The rest of the consensus algorithms in this paper have been systematically covered to address the limitations of proof-of-work. This paper also covered Raft and PBFT consensus algorithms suitable for permissioned networks. Although the PBFT consensus protocol has a high throughput and a low latency, it has limited node scalability. The PBFT has a low byzantine fault tolerant rate. This paper also covers PoEWAL for blockchain-based IoT applications and WBFT, which prevents corrupt nodes from taking part in consensus. A comparative analysis of the consensus algorithms provides an explicit knowledge of the present research, which also offers guidance for future study. © 2024

关键词： Blockchain technology Byzantine faults Consensus algorithm Latency Throughput

来源：评论

学校读者我要写书评

暂无评论

Konjac glucomannan (KGM) aerogel immobilized microalgae: A new way for marine oil spills remediation

引用

Marine Pollution Bulletin 2025年 218卷 118158页

作者： Wang, Xudong Du, Zhuorong Song, Zhao Liu, Mei He, Pengfei Feng, Lijuan Chen, Zhi Chen, Qingguo Zhejiang Key Laboratory of Pollution Control for Port-Petrochemical Industry Zhejiang Ocean University Zhoushan316022 China National & local Joint Engineering Research Center of Harbor Oil & Gas Storage and Transportation Technology Zhejiang Ocean University Zhoushan316022 China School of Marine Science & Technology Zhejiang Ocean University Zhoushan316022 China Zhejiang Marine Fisheries Research Institute Zhoushan316021 China Department of Building Civil and Environmental Engineering Faculty of Engineering & Computer Sciences Concordia University MontrealQCH3G1M8 Canada

This study investigated the oil pollution remediation ability of konjac glucomannan (KGM) aerogel immobilized Chlorella vulgaris LH-1 (C. vulgaris). The KGM aerogel exhibited a high porosity of 86.83 %, and with the change of crude oil concentration (1–10 g/L), the crude oil adsorption efficiency of KGM aerogel was up to 18.75 %. The KGM aerogel carrier enhanced the crude oil biodegradation efficiency of C. vulgaris by 13.86 %. The concentration of residual n-alkanes of the immobilized C. vulgaris group was only 36.83 % of that of the free C. vulgaris group after 14 days of cultivation. In the crude oil metabolites of immobilized C. vulgaris, the concentrations of benzoic acid, phthalic acid, and benzophenone were 1.35, 21.18 and 3.43 times higher than those of the free C. vulgaris, respectively. Compared with free C. vulgaris group, 9 monooxygenase genes, 11 dioxygenase genes and 2 cytochrome P450 genes of immobilized C. vulgaris group were up-regulation and the "hydrocarbon metabolism process" was significantly enriched (enrichment factor is 0.85) in the gene ontology (GO) analysis of the transcriptome of C. vulgaris. These results showed that the ability of C. vulgaris to degrade crude oil was improved due to the existence of the KGM aerogel immobilized carrier. © 2025 Elsevier Ltd

关键词： Marine pollution

来源：评论

学校读者我要写书评

暂无评论

Scalability and efficiency challenges for the exascale supercomputing system:practice of a parallel supporting environment on the Sunway exascale prototype system

引用

Frontiers of Information technology & Electronic engineering 2023年第1期24卷 41-58页

作者： Xiaobin HE Xin CHEN Heng GUO Xin LIU Dexun CHEN Yuling YANG Jie GAO Yunlong FENG Longde CHEN Xiaona DIAO Zuoning CHEN National Research Center of Parallel Computer Engineering and Technology Beijing 100190China

With the continuous improvement of supercomputer performance and the integration of artificial intelligence with traditional scientific computing,the scale of applications is gradually increasing,from millions to tens of millions of computing cores,which raises great challenges to achieve high scalability and efficiency of parallel applications on super-large-scale *** the Sunway exascale prototype system as an example,in this paper we first analyze the challenges of high scalability and high efficiency for parallel applications in the exascale *** overcome these challenges,the optimization technologies used in the parallel supporting environment software on the Sunway exascale prototype system are highlighted,including the parallel operating system,input/output(I/O)optimization technology,ultra-large-scale parallel debugging technology,10-million-core parallel algorithm,and mixed-precision *** operating systems and I/O optimization technology mainly support largescale system scaling,while the ultra-large-scale parallel debugging technology,10-million-core parallel algorithm,and mixed-precision method mainly enhance the efficiency of large-scale ***,the contributions to various applications running on the Sunway exascale prototype system are introduced,verifying the effectiveness of the parallel supporting environment design.

关键词： parallel computing Sunway Ultra-large-scale Supercomputer

来源：评论

学校读者我要写书评

暂无评论

Integrative Multi-Source Mapping of Coastal Environments Using Hy-perspectral Data Fusion Platforms

引用

IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 2025年

作者： Han, Huimin Por, Lip Yee Talpur, Bandeh Ali Hashim, Muhammad Yang, Jing Saidani, Taoufik Djumaniyazova, Mukhayya Xusinovna Madrakhimovich, Khudoynazarov Egambergan Talpur, Kazim Raza Hainan Vocational University of Science and Technology Mechanical and Electrical Engineering Col-lege Haikou571126 China Uni-versiti Malaya Department of Computer System and Technology Kuala Lumpur50603 Malaysia Trinity Col-lege Dublin School of Computer Science and Statistics Ireland National Textile Uni-versity Faisalabad Faisalabad Business School Pakistan Universiti Malaya Department of Computer System and Technology Faculty of Computer Science and Information Technology Kuala Lumpur50603 Malaysia Northern Border University Center for Scientific Research and Entrepreneurship Arar73213 Saudi Arabia Urganch State University Pedagogy and Psychology Department Uzbekistan Mamun University Uzbekistan Jamshoro Institute of Health Management and Re-search Sciences Sindh Pakistan

Hyperspectral images (HSIs) not only have a large number of spectral features, but also show a comprehensive spa-tial distribution of land cover and offer substantial advantages in the fine classification of ground materials. However, due to the high dimensionality and redundancy of HSIs and class imbalance in hyperspectral datasets, improving classification performance is still a huge challenge. Recently, graph-based learning algorithms are proven to be better classifiers for high dimensional data. Learning graph data and modeling spatial topological links be-tween features are significantly more accessible and more effective with graph-based deep learning, which has been demonstrated to be a promising approach. In the past few years, convolutional neu-ral networks (CNNs) and graph convolutional networks (GCNs) have achieved good results in HSI classification, but CNNs strug-gle to achieve good accuracy in low samples, while GCNs have a substantial computational cost. To resolve these issues, this paper proposes a feature fusion model using a 3D-CNN and hybrid-GAT named FFCHG. The algorithm consists of two elements: the 3D-CNN, which produces good classification for 3D HSI cube data, and GAT-based encoder and decoder modules that help in im-proving the classification accuracy of the 3D-CNN. Finally, the multiple features are merged with the help of two neural network models. Experiments on three public HSI datasets show that the proposed methods perform better than other state-of-the-art methods using the limited training samples and in low classifica-tion time with accuracy of more than 90% in all datasets and kappa with more than 91% for each dataset. © 2025 IEEE.

关键词： Convolutional neural networks

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：