检索结果-内蒙古大学图书馆

SSRN 2024年

作者： Alhaek, Fares Li, Tianrui Rajeh, Taha M. Javed, Muhammad Hafeez Liang, Weichao School of Computing and Artificial Intelligence Southwest Jiaotong University Chengdu611756 China Engineering Research Center of Sustainable Urban Intelligent Transportation Ministry of Education Chengdu611756 China National Engineering Laboratory of Integrated Transportation Big Data Application Technology Southwest Jiaotong University Chengdu611756 China Manufacturing Industry Chains Collaboration and Information Support Technology Key Laboratory of Sichuan Province Southwest Jiaotong University Chengdu611756 China

The proliferation of vehicles and the intricate layout of road systems have contributed to a significant rise in traffic accidents, posing a pressing concern globally. Despite the advancements facilitated by deep learning, several challenges persist in the domain of traffic accident prediction. Firstly, the sparsity of accident data in certain city regions presents a significant obstacle, particularly when attempting fine-grained predictions at a local level. Secondly, the intricate spatial-temporal relations inherent in traffic accident data pose a challenge for existing prediction models. In response to these formidable challenges, this paper presents GLST-TARP, a novel model for predicting traffic accident risk in urban environments by leveraging both global semantic and localized geographic spatial-temporal relations. Specifically, we construct multi-graphs to encode global static and dynamic spatial relations, incorporating attention mechanisms to adaptively focus on relevant information and temporal relations. Additionally, we employ channel-wise convolutional neural network blocks to extract localized geographic features and enhance predictive accuracy. The proposed model is trained using a Huber loss function tailored for regression tasks, relieving the impact of zero values during optimization. Experimental results demonstrate that GLST-TARP outperforms state-of-the-art methods in predicting traffic accident risks, showcasing its potential for enhancing urban safety and transportation management. © 2024, The Authors. All rights reserved.

关键词： Graph neural networks

来源：评论

学校读者我要写书评

暂无评论

Privacy-Preserving Federated Learning Framework with Resisting Source Inference Attacks in Smart Healthcare

SSRN

引用

SSRN 2024年

作者： Chenghe, Dong Xu, Guangquan Zhang, Jianhong Bai, Hongpeng Hao, Peng College of Intelligence and Computing Tianjin University Tianjin300354 China School of Big Data Qingdao Huanghai University Qingdao China Tianjin300350 China Department of Electrical and Computer Engineering North China University of Technology Beijing100144 China Key Laboratory of Intelligent Education Technology Application of Zhejiang Province School of Computer Science and Technology Zhejiang Normal University Jinhua321004 China

Federated Learning (FL) is a distributed machine learning technology that is extensive applications in smart healthcare, since it allows multiple medical institutions to jointly train medical diagnostic models without sharing their local sensitive dataset with others. Despite its benefits, FL is susceptible to source inference attacks and model poisoning. To avoid these issues, we propose a federated learning framework for smart healthcare that preserves privacy and employs certificateless ring signatures to counteract potential attacks. During the collaborative training process of the FL model, certificateless ring signatures are used to obscure the origin of parameter updates, ensuring the anonymity of the parameter sources. This approach has the advantage of being computationally efficient due to no complicated pairing operations. Furthermore, compared with the previous certificateless ring signature schemes, our proposed scheme is more effective in reducing the success rate of source inference attacks, and is efficient in terms of computation and communication overhead. © 2024, The Authors. All rights reserved.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

DNN Task Assignment in UAV Networks: A Generative AI Enhanced Multi-Agent Reinforcement Learning Approach

arXiv

引用

arXiv 2024年

作者： Tang, Xin Chen, Qian Weng, Wenjie Liao, Binhan Wang, Jiacheng Cao, Xianbin Li, Xiaohuan Guangxi University Key Laboratory of Intelligent Networking and Scenario System School of Information and Communication Guilin University of Electronic Technology Guilin541004 China The College of Computing and Data Science Nanyang Technological University 639798 Singapore School of Architecture and Transportation Engineering Guilin University of Electronic Technology Guilin541004 China National Engineering Laboratory for Comprehensive Transportation Big Data Application Technology Guangxi Nanning530001 China College of Computing and Data Science Nanyang Technological University 639798 Singapore School of Electronic and Information Engineering The Key Laboratory of Advanced Technology of Near Space Information System Ministry of Industry and Information Technology of China Beihang University Beijing100191 China

Unmanned Aerial Vehicles (UAVs) possess high mobility and flexible deployment capabilities, prompting the development of UAVs for various application scenarios within the Internet of Things (IoT). The unique capabilities of UAVs give rise to increasingly critical and complex tasks in uncertain and potentially harsh environments. The substantial amount of data generated from these applications necessitates processing and analysis through deep neural networks (DNNs). However, UAVs encounter challenges due to their limited computing resources when managing DNN models. This paper presents a joint approach that combines multiple-agent reinforcement learning (MARL) and generative diffusion models (GDM) for assigning DNN tasks to a UAV swarm, aimed at reducing latency from task capture to result output. To address these challenges, we first consider the task size of the target area to be inspected and the shortest flying path as optimization constraints, employing a greedy algorithm to resolve the subproblem with a focus on minimizing the UAV’s flying path and the overall system cost. In the second stage, we introduce a novel DNN task assignment algorithm, termed GDM-MADDPG, which utilizes the reverse denoising process of GDM to replace the actor network in multi-agent deep deterministic policy gradient (MADDPG). This approach generates specific DNN task assignment actions based on agents’ observations in a dynamic environment. Simulation results indicate that our algorithm performs favorably compared to benchmarks in terms of path planning, Age of Information (AoI), energy consumption, and task load balancing. Copyright © 2024, The Authors. All rights reserved.

关键词： Motion planning

来源：评论

学校读者我要写书评

暂无评论

A Diverse Biases Non-negative Latent Factorization of Tensors Model for Dynamic Network Link Prediction

A Diverse Biases Non-negative Latent Factorization of Tensor...

引用

IEEE International Conference on Networking, Sensing and Control

作者： Xuke Wu Hang Gou Hao Wu Juan Wang Minzhi Chen Siyu Lai Computer School China West Normal University Nanchong China Computer School of China West Normal University Nanchong Sichuan China University of Chinese Academy of Sciences Beijing China Chongqing Engineering Research Center of Big Data Application for Smart Cities and Chongqing Key Laboratory of Big Data and Intelligent Computing Chongqing Institute of Green and Intelligent Technology Chinese Academy of Sciences Chongqing China North Sichuan Medical College Nanchong China Medical College Nanchong Sichuan China

ISBN: (数字)9781728168555

ISBN: (纸本)9781728168562

Dynamic networks vary over time, making it vital to capture networks temporal patterns for predicting missing links with high accuracy. A biased non-negative latent factorization of tensors (BNLFT) model is very effective in extracting such patterns from dynamic data. However, a BNLFT model only integrates single bias, which cannot adequately represents the volatility of the dynamic data. To address this issue, this paper presents a Diverse Biases Non-negative Latent Factorization of Tensors (DBNT) model for accurate prediction of missing links in dynamic networks. Meanwhile, for further prediction accuracy improvement, the preprocessing bias is integrated into the DBNT model. Empirical studies on two dynamic networks datasets from real applications show that compared with state of the art predictors, a DBNT model achieves higher prediction accuracy.

关键词： Tensors Conferences Predictive models Data models Sensors Data mining

来源：评论

学校读者我要写书评

暂无评论

Layer-Wise Feature Metric of Semantic-Pixel Matching for Few-Shot Learning

arXiv

引用

arXiv 2024年

作者： Tang, Hao Lu, Junhao Huang, Guoheng Li, Ming Chen, Xuhang Zhong, Guo Tan, Zhengguang Li, Zinuo School of Computer Science and Technology Guangdong University of Technology Guangdong China School of Computer Science and Engineering Huizhou University Huizhou China Zhejiang Institute of Optoelectronics Jinhua China Zhejiang Key Laboratory of Intelligent Education Technology and Application Jinhua China School of Information Science and Technology Guangdong University of Foreign Studies Guangzhou China School of Physics Maths and Computing University of Western Australia WA Australia

In Few-Shot Learning (FSL), traditional metric-based approaches often rely on global metrics to compute similarity. However, in natural scenes, the spatial arrangement of key instances is often inconsistent across images. This spatial misalignment can result in mismatched semantic pixels, leading to inaccurate similarity measurements. To address this issue, we propose a novel method called the Layer-Wise Features Metric of Semantic-Pixel Matching (LWFM-SPM) to make finer comparisons. Our method enhances model performance through two key modules: (1) the Layer-Wise Embedding (LWE) Module, which refines the cross-correlation of image pairs to generate well-focused feature maps for each layer;(2)the Semantic-Pixel Matching (SPM) Module, which aligns critical pixels based on semantic embeddings using an assignment algorithm. We conducted extensive experiments to evaluate our method on four widely used few-shot classification benchmarks: miniImageNet, tieredImageNet, CUB-200-2011, and CIFAR-FS. The results indicate that LWFM-SPM achieves competitive performance across these benchmarks. Our code will be publicly available on https://***/Halo2Tang/Code-for-LWFM-SPM. Copyright © 2024, The Authors. All rights reserved.

关键词： Zero-shot learning

来源：评论

学校读者我要写书评

暂无评论

Predicting Large-scale Protein-protein Interactions by Extracting Coevolutionary Patterns with MapReduce Paradigm

Predicting Large-scale Protein-protein Interactions by Extra...

引用

IEEE International Conference on Systems, Man and Cybernetics

作者： Lun Hu Bo-Wei Zhao Shicheng Yang Xin Luo MengChu Zhou Xinjiang Technical Institute of Physics and Chemistry Chinese Academy of Sciences Urumqi China School of Computer Science and Technology Wuhan University of Technology Wuhan China Chongqing Engineering Research Center of Big Data Application for Smart Cities and Chongqing Key Laboratory of Big Data and Intelligent Computing Chongqing Institute of Green and Intelligent Technology Chinese Academy of Sciences Chongqing China Hengrui (Chongqing) Artificial Intelligence Research Center Cloudwalk China New Jersey Institute of Technology Newark NJ USA

ISBN: (纸本)9781665442084

Protein-protein interactions are of great significance for us to understand the functional mechanisms of proteins. With the rapid development of high-throughput genomic technology, the amount of protein-protein interaction data has become so big that most of existing prediction algorithms are no longer applicable. To address this problem, we develop a distributed framework by reimplementing one of state-of-the-art algorithms, i.e., CoFex, by using MapReduce. In particular, we adopt a novel tree-based data structure to reduce the heavy memory consumption cased by the huge sequence information of proteins. After that, the procedure of CoFex is modified by following the paradigm of MapReduce such that the prediction task can be completed in a distributed manner, thus fulfilling the demanding requirements of large-scale protein-protein interaction prediction. A series of experiments have been conducted to evaluate the performance of the proposed distributed framework in terms of both efficiency and effectiveness. Experimental results demonstrate that the proposed framework can considerably improve the efficiency of CoFex by achieving more than two-orders-of-magnitude improvement in computational efficiency while retaining a comparable level of accuracy.

关键词： Proteins Memory management Genomics Predictive models Prediction algorithms Data structures Computational efficiency

来源：评论

学校读者我要写书评

暂无评论

Blockchain-based lightweight trusted data interaction scheme for cross-domain IIoT

引用

Digital Communications and Networks 2024年

作者： Fengqi Li Yingjie Zhao Kaiyang Zhang Hui Xu Yanjuan Wang Deguang Wang School of Railway Intelligent Engineering Dalian Jiaotong University Dalian 116028 China Dalian Key Laboratory of Blockchain Technology and Application Dalian 116028 China Yunnan Key Laboratory of Service Computing Kunming 650221 China

To facilitate cross-domain data interaction in the Industrial Internet of Things (IIoT), establishing trust between multiple administrative domains is essential. Although blockchain technology has been proposed as a solution, current techniques still suffer from issues related to efficiency, security, and privacy. Our research aims to address these challenges by proposing a lightweight, trusted data interaction scheme based on blockchain, which reduces redundant interactions among entities. We enhance the traditional Practical Byzantine Fault Tolerance (PBFT) algorithm to support lightweight distributed consensus in large-scale IIoT scenarios. Introducing a composite digital signature algorithm and incorporating veto power minimizes resource consumption and eliminates ineffective consensus operations. The experimental results show that, compared with PBFT, our scheme reduces latency by 27.2%, thereby improving communication efficiency and resource utilization. Furthermore, we develop a lightweight authentication technique specifically for cross-domain IIoT, leveraging blockchain technology to achieve distributed collaborative authentication. The performance comparisons indicate that our method significantly outperforms traditional schemes, with an average authentication latency of approximately 151 milliseconds. Additionally, we introduce a trusted federated learning (FL) algorithm that ensures comprehensive trust assessments for devices across different domains while protecting data privacy. Extensive simulations and experiments validate the reliability of our approach.

关键词： Blockchain Industrial internet of things Federated learning Cross-domain Authentication

来源：评论

学校读者我要写书评

暂无评论

How Universal Polynomial Bases Enhance Spectral Graph Neural Networks: Heterophily, Over-smoothing, and Over-squashing

arXiv

引用

arXiv 2024年

作者： Huang, Keke Wang, Yu Guang Li, Ming Liò, Pietro School of Computing National University of Singapore Singapore Institute of Natural Sciences School of Mathematical Sciences Zhangjiang Institute for Advanced Study Shanghai Jiao Tong University Shanghai China Shanghai AI Laboratory Shanghai China School of Mathematics and Statistics University of New South Wales Sydney Australia Zhejiang Institute of Optoelectronics Jinhua China Zhejiang Key Laboratory of Intelligent Education Technology and Application Zhejiang Normal University Jinhua China Department of Computer Science and Technology Cambridge University Cambridge United Kingdom

Spectral Graph Neural Networks (GNNs), alternatively known as graph filters, have gained increasing prevalence for heterophily graphs. Optimal graph filters rely on Laplacian eigendecomposition for Fourier transform. In an attempt to avert prohibitive computations, numerous polynomial filters have been proposed. However, polynomials in the majority of these filters are predefined and remain fixed across different graphs, failing to accommodate the varying degrees of heterophily. Addressing this gap, we demystify the intrinsic correlation between the spectral property of desired polynomial bases and the heterophily degrees via thorough theoretical analyses. Subsequently, we develop a novel adaptive heterophily basis wherein the basis vectors mutually form angles reflecting the heterophily degree of the graph. We integrate this heterophily basis with the homophily basis to construct a universal polynomial basis UniBasis, which devises a polynomial filter-based graph neural network – UniFilter. It optimizes the convolution and propagation in GNN, thus effectively limiting over-smoothing and alleviating over-squashing. Our extensive experiments, conducted on datasets with a diverse range of heterophily, support the superiority of UniBasis in the universality but also its proficiency in graph explanation. © 2024, CC BY-NC-SA.

关键词： Graph neural networks

来源：评论

学校读者我要写书评

暂无评论

How universal polynomial bases enhance spectral graph neural networks: heterophily, over-smoothing, and over-squashing 24

How universal polynomial bases enhance spectral graph neural...

引用

Proceedings of the 41st International Conference on Machine Learning

作者： Keke Huang Yu Guang Wang Ming Li Pietro Liò School of Computing National University of Singapore Singapore Institute of Natural Sciences School of Mathematical Sciences Zhangjiang Institute for Advanced Study Shanghai Jiao Tong University Shanghai China and Shanghai AI Laboratory Shanghai China and School of Mathematics and Statistics University of New South Wales Sydney Australia Zhejiang Institute of Optoelectronics Jinhua China and Zhejiang Key Laboratory of Intelligent Education Technology and Application Zhejiang Normal University Jinhua China Department of Computer Science and Technology Cambridge University Cambridge UK

Spectral Graph Neural Networks (GNNs), alternatively known as graph filters, have gained increasing prevalence for heterophily graphs. Optimal graph filters rely on Laplacian eigendecomposition for Fourier transform. In an attempt to avert prohibitive computations, numerous polynomial filters have been proposed. However, polynomials in the majority of these filters are predefined and remain fixed across different graphs, failing to accommodate the varying degrees of heterophily. Addressing this gap, we demystify the intrinsic correlation between the spectral property of desired polynomial bases and the heterophily degrees via thorough theoretical analyses. Subsequently, we develop a novel adaptive heterophily basis wherein the basis vectors mutually form angles reflecting the heterophily degree of the graph. We integrate this heterophily basis with the homophily basis to construct a universal polynomial basis UniBasis, which devises a polynomial filter-based graph neural network - UniFilter. It optimizes the convolution and propagation in GNN, thus effectively limiting over-smoothing and alleviating over-squashing. Our extensive experiments, conducted on datasets with a diverse range of heterophily, support the superiority of UniBasis in the universality but also its proficiency in graph explanation.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Towards Graph Self-Supervised Learning with Contrastive Adjusted Zooming

arXiv

引用

arXiv 2021年

作者： Zheng, Yizhen Jin, Ming Pan, Shirui Li, Yuan-Fang Peng, Hao Li, Ming Li, Zhao Department of Data Science and AI Faculty of IT Monash University ClaytonVIC3800 Australia School of Information and Communication Technology Griffith University Australia SouthportQLD4222 Australia Beijing Advanced Innovation Center for Big Data and Brain Computing Beihang University Beijing100191 China Key Laboratory of Intelligent Education Technology Application of Zhejiang Province Zhejiang Normal University Jinhua321004 China Alibaba Group Hangzhou311100 China

Graph representation learning (GRL) is critical for graph-structured data analysis. However, most of the existing graph neural networks (GNNs) heavily rely on labeling information, which is normally expensive to obtain in the real world. Although some existing works aim to effectively learn graph representations in an unsupervised manner, they suffer from certain limitations, such as the heavy reliance on monotone contrastiveness and limited scalability. To overcome the aforementioned problems, in light of the recent advancements in graph contrastive learning, we introduce a novel self-supervised graph representation learning algorithm via Graph Contrastive Adjusted Zooming, namely G-Zoom, to learn node representations by leveraging the proposed adjusted zooming scheme. Specifically, this mechanism enables G-Zoom to explore and extract self-supervision signals from a graph from multiple scales: micro (i.e., node-level), meso (i.e., neighborhood-level), and macro (i.e., subgraph-level). Firstly, we generate two augmented views of the input graph via two different graph augmentations. Then, we establish three different contrastiveness on the above three scales progressively, from node, neighboring, to subgraph level, where we maximize the agreement between graph representations across scales. While we can extract valuable clues from a given graph on the micro and macro perspectives, the neighboring-level contrastiveness offers G-Zoom the capability of a customizable option based on our adjusted zooming scheme to manually choose an optimal viewpoint that lies between the micro and macro perspectives to better understand the graph data. Additionally, to make our model scalable to large graphs, we employ a parallel graph diffusion approach to decouple model training from the graph size. We have conducted extensive experiments on real-world datasets, and the results demonstrate that our proposed model outperforms state-of-the-art methods consistently. © 2021, CC BY.

关键词： Graph 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：