随着慢性病成为全球公共卫生领域的核心挑战,患者对高质量、个性化、可持续的健康信息服务需求日益增长。近年来,尽管大语言模型在医疗问答任务中展现出显著优势,但现有系统在面对慢性病管理这一长期性、情境化、语义复杂的场景时,仍面...随着慢性病成为全球公共卫生领域的核心挑战,患者对高质量、个性化、可持续的健康信息服务需求日益增长。近年来,尽管大语言模型在医疗问答任务中展现出显著优势,但现有系统在面对慢性病管理这一长期性、情境化、语义复杂的场景时,仍面临知识调度粒度粗、语义感知能力弱与模型响应一致性差等瓶颈。为此,本研究遵循“结构化知识建模—混合驱动检索—大模型协同应答”三阶段流程,提出一种融合图检索增强生成(graph retrievalaugmented generation,GraphRAG)与混合专家网络(mixture of experts,MoE)的慢性病问答系统。首先,构建覆盖疾病演化、生活方式干预和长期管理要素的多源异构专有知识图谱,通过结合关键词匹配、语义向量召回与多子图提示融合的三通道混合图谱检索机制,实现对复杂信息需求的动态知识调度;其次,初筛用户查询合法性与规范性,结合用户隐含意图与融合子图,通过MoE动态门控智能体对9种开源大模型进行指令级协同,实现语义深度与生成精度的双擎增强;最后,引入常识校验、内容完整性检查与格式化规则,精细预生成文本,确保输出结果的准确性、完整性与语言可读性。本研究自建覆盖118种慢性病与3类管理任务的测试集CdMedQA,综合采用客观指标评价和主观满意度对比两种方式验证系统性能。结果表明,本研究构建系统在准确性、清晰度、个性化及情境适应性等方面显著优于多种通用与医疗垂直大模型基线。研究结果不仅为慢性病智能化管理提供了新路径,也为多源知识驱动下人机交互优化与生成内容可信性提升提供了理论支持与技术方案。展开更多
基于深度学习的网络攻击检测是对欧几里得数据进行建模,无法学习攻击数据中的结构特征。为此,提出一种基于改进图采样与聚合(graph sample and aggregate,GraphSAGE)的网络攻击检测算法。首先,将攻击数据从平面结构转换为图结构数据。其...基于深度学习的网络攻击检测是对欧几里得数据进行建模,无法学习攻击数据中的结构特征。为此,提出一种基于改进图采样与聚合(graph sample and aggregate,GraphSAGE)的网络攻击检测算法。首先,将攻击数据从平面结构转换为图结构数据。其次,对GraphSAGE算法进行了改进,包括在消息传递阶段融合节点和边的特征,同时在消息聚合过程中考虑不同源节点对目标节点的影响程度,并在边嵌入生成时引入残差学习机制。在两个公开网络攻击数据集上的实验结果表明,在二分类情况下,所提算法的总体性能优于E-GraphSAGE、LSTM、RNN、CNN算法;在多分类情况下,所提算法在大多数攻击类型上的F1值高于对比算法。展开更多
Fault diagnosis occupies a pivotal position within the domain of machine and equipment management.Existing methods,however,often exhibit limitations in their scope of application,typically focusing on specific types o...Fault diagnosis occupies a pivotal position within the domain of machine and equipment management.Existing methods,however,often exhibit limitations in their scope of application,typically focusing on specific types of signals or faults in individual mechanical components while being constrained by data types and inherent characteristics.To address the limitations of existing methods,we propose a fault diagnosis method based on graph neural networks(GNNs)embedded with multirelationships of intrinsic mode functions(MIMF).The approach introduces a novel graph topological structure constructed from the features of intrinsic mode functions(IMFs)of monitored signals and their multirelationships.Additionally,a graph-level based fault diagnosis network model is designed to enhance feature learning capabilities for graph samples and enable flexible application across diverse signal sources and devices.Experimental validation with datasets including independent vibration signals for gear fault detection,mixed vibration signals for concurrent gear and bearing faults,and pressure signals for hydraulic cylinder leakage characterization demonstrates the model's adaptability and superior diagnostic accuracy across various types of signals and mechanical systems.展开更多
Urban air pollution has brought great troubles to physical and mental health,economic development,environmental protection,and other aspects.Predicting the changes and trends of air pollution can provide a scientific ...Urban air pollution has brought great troubles to physical and mental health,economic development,environmental protection,and other aspects.Predicting the changes and trends of air pollution can provide a scientific basis for governance and prevention efforts.In this paper,we propose an interval prediction method that considers the spatio-temporal characteristic information of PM_(2.5)signals from multiple stations.K-nearest neighbor(KNN)algorithm interpolates the lost signals in the process of collection,transmission,and storage to ensure the continuity of data.Graph generative network(GGN)is used to process time-series meteorological data with complex structures.The graph U-Nets framework is introduced into the GGN model to enhance its controllability to the graph generation process,which is beneficial to improve the efficiency and robustness of the model.In addition,sparse Bayesian regression is incorporated to improve the dimensional disaster defect of traditional kernel density estimation(KDE)interval prediction.With the support of sparse strategy,sparse Bayesian regression kernel density estimation(SBR-KDE)is very efficient in processing high-dimensional large-scale data.The PM_(2.5)data of spring,summer,autumn,and winter from 34 air quality monitoring sites in Beijing verified the accuracy,generalization,and superiority of the proposed model in interval prediction.展开更多
Tracking and analyzing data from research projects is critical for understanding research trends and supporting the development of science and technology strategies.However,the data from these projects is often comple...Tracking and analyzing data from research projects is critical for understanding research trends and supporting the development of science and technology strategies.However,the data from these projects is often complex and inadequate,making it challenging for researchers to conduct in-depth data mining to improve policies or management.To address this problem,this paper adopts a top-down approach to construct a knowledge graph(KG)for research projects.Firstly,we construct an integrated ontology by referring to the metamodel of various architectures,which is called the meta-model integration conceptual reference model.Subsequently,we use the dependency parsing method to extract knowledge from unstructured textual data and use the entity alignment method based on weakly supervised learning to classify the extracted entities,completing the construction of the KG for the research projects.In addition,a knowledge inference model based on representation learning is employed to achieve knowledge completion and improve the KG.Finally,experiments are conducted on the KG for research projects and the results demonstrate the effectiveness of the proposed method in enriching incomplete data within the KG.展开更多
Given two graphs G and H,the Ramsey number R(G,H)is the smallest positive integer N such that every 2-coloring of the edges of K_(N)contains either a red G or a blue H.Let K_(N-1)■K_(1,k)be the graph obtained from K_...Given two graphs G and H,the Ramsey number R(G,H)is the smallest positive integer N such that every 2-coloring of the edges of K_(N)contains either a red G or a blue H.Let K_(N-1)■K_(1,k)be the graph obtained from K_(N-1)by adding anew vertexνconnecting k vertices of K_(N-1).A graph G withχ(G)=k+1 is called edge-critical if G contains an edge e such thatχ(G-e)=k.A considerable amount of research has been conducted by previous scholars on Ramsey numbers ofgraphs.In this study,we show that for an edge-critical graph G with x(G)=k+1,when k≥2,1≥2,and n is sufficiently large,R(G,K_(1)+nK_(t))=knt+1 and r,(G,K_(1)+nK_(t))=(k-1)nt+1.展开更多
A nowhere-zero k-flow on a graph G=(V(G),E(G))is a pair(D,f),where D is an orientation on E(G)and f:E(G)→{±1,±2,,±(k-1)}is a function such that the total outflow equals to the total inflow at each vert...A nowhere-zero k-flow on a graph G=(V(G),E(G))is a pair(D,f),where D is an orientation on E(G)and f:E(G)→{±1,±2,,±(k-1)}is a function such that the total outflow equals to the total inflow at each vertex.This concept was introduced by Tutte as an extension of face colorings,and Tutte in 1954 conjectured that every bridgeless graph admits a nowhere-zero 5-flow,known as the 5-Flow Conjecture.This conjecture is verified for some graph classes and remains unresolved as of today.In this paper,we show that every bridgeless graph of Euler genus at most 20 admits a nowhere-zero 5-flow,which improves several known results.展开更多
Urban combat environments pose complex and variable challenges for UAV path planning due to multidimensional factors,such as static and dynamic obstructions as well as risks of exposure to enemy detection,which threat...Urban combat environments pose complex and variable challenges for UAV path planning due to multidimensional factors,such as static and dynamic obstructions as well as risks of exposure to enemy detection,which threaten flight safety and mission success.Traditional path planning methods typically depend solely on the distribution of static obstacles to generate collision-free paths,without accounting for constraints imposed by enemy detection and strike capabilities.Such a simplified approach can yield safety-compromising routes in highly complex urban airspace.To address these limitations,this study proposes a multi-parameter path planning method based on reachable airspace visibility graphs,which integrates UAV performance constraints,environmental limitations,and exposure risks.An innovative heuristic algorithm is developed to balance operational safety and efficiency by both exposure risks and path length.In the case study set in a typical mixed-use urban area,analysis of airspace visibility graphs reveals significant variations in exposure risk at different regions and altitudes due to building encroachments.Path optimization results indicate that the method can effectively generate covert and efficient flight paths by dynamically adjusting the exposure index,which represents the likelihood of enemy detection,and the path length,which corresponds to mission execution time.展开更多
For target tracking and localization in bearing-only sensor network,it is an essential and significant challenge to solve the problem of plug-and-play expansion while stably enhancing the accuracy of state estimation....For target tracking and localization in bearing-only sensor network,it is an essential and significant challenge to solve the problem of plug-and-play expansion while stably enhancing the accuracy of state estimation.This paper pro-poses a distributed state estimation method based on two-layer factor graph.Firstly,the measurement model of the bearing-only sensor network is constructed,and by investigating the observ-ability and the Cramer-Rao lower bound of the system model,the preconditions are analyzed.Subsequently,the location fac-tor graph and cubature information filtering algorithm of sensor node pairs are proposed for localized estimation.Building upon this foundation,the mechanism for propagating confidence mes-sages within the fusion factor graph is designed,and is extended to the entire sensor network to achieve global state estimation.Finally,groups of simulation experiments are con-ducted to compare and analyze the results,which verifies the rationality,effectiveness,and superiority of the proposed method.展开更多
文摘随着慢性病成为全球公共卫生领域的核心挑战,患者对高质量、个性化、可持续的健康信息服务需求日益增长。近年来,尽管大语言模型在医疗问答任务中展现出显著优势,但现有系统在面对慢性病管理这一长期性、情境化、语义复杂的场景时,仍面临知识调度粒度粗、语义感知能力弱与模型响应一致性差等瓶颈。为此,本研究遵循“结构化知识建模—混合驱动检索—大模型协同应答”三阶段流程,提出一种融合图检索增强生成(graph retrievalaugmented generation,GraphRAG)与混合专家网络(mixture of experts,MoE)的慢性病问答系统。首先,构建覆盖疾病演化、生活方式干预和长期管理要素的多源异构专有知识图谱,通过结合关键词匹配、语义向量召回与多子图提示融合的三通道混合图谱检索机制,实现对复杂信息需求的动态知识调度;其次,初筛用户查询合法性与规范性,结合用户隐含意图与融合子图,通过MoE动态门控智能体对9种开源大模型进行指令级协同,实现语义深度与生成精度的双擎增强;最后,引入常识校验、内容完整性检查与格式化规则,精细预生成文本,确保输出结果的准确性、完整性与语言可读性。本研究自建覆盖118种慢性病与3类管理任务的测试集CdMedQA,综合采用客观指标评价和主观满意度对比两种方式验证系统性能。结果表明,本研究构建系统在准确性、清晰度、个性化及情境适应性等方面显著优于多种通用与医疗垂直大模型基线。研究结果不仅为慢性病智能化管理提供了新路径,也为多源知识驱动下人机交互优化与生成内容可信性提升提供了理论支持与技术方案。
文摘基于深度学习的网络攻击检测是对欧几里得数据进行建模,无法学习攻击数据中的结构特征。为此,提出一种基于改进图采样与聚合(graph sample and aggregate,GraphSAGE)的网络攻击检测算法。首先,将攻击数据从平面结构转换为图结构数据。其次,对GraphSAGE算法进行了改进,包括在消息传递阶段融合节点和边的特征,同时在消息聚合过程中考虑不同源节点对目标节点的影响程度,并在边嵌入生成时引入残差学习机制。在两个公开网络攻击数据集上的实验结果表明,在二分类情况下,所提算法的总体性能优于E-GraphSAGE、LSTM、RNN、CNN算法;在多分类情况下,所提算法在大多数攻击类型上的F1值高于对比算法。
文摘Fault diagnosis occupies a pivotal position within the domain of machine and equipment management.Existing methods,however,often exhibit limitations in their scope of application,typically focusing on specific types of signals or faults in individual mechanical components while being constrained by data types and inherent characteristics.To address the limitations of existing methods,we propose a fault diagnosis method based on graph neural networks(GNNs)embedded with multirelationships of intrinsic mode functions(MIMF).The approach introduces a novel graph topological structure constructed from the features of intrinsic mode functions(IMFs)of monitored signals and their multirelationships.Additionally,a graph-level based fault diagnosis network model is designed to enhance feature learning capabilities for graph samples and enable flexible application across diverse signal sources and devices.Experimental validation with datasets including independent vibration signals for gear fault detection,mixed vibration signals for concurrent gear and bearing faults,and pressure signals for hydraulic cylinder leakage characterization demonstrates the model's adaptability and superior diagnostic accuracy across various types of signals and mechanical systems.
基金Project(2020YFC2008605)supported by the National Key Research and Development Project of ChinaProject(52072412)supported by the National Natural Science Foundation of ChinaProject(2021JJ30359)supported by the Natural Science Foundation of Hunan Province,China。
文摘Urban air pollution has brought great troubles to physical and mental health,economic development,environmental protection,and other aspects.Predicting the changes and trends of air pollution can provide a scientific basis for governance and prevention efforts.In this paper,we propose an interval prediction method that considers the spatio-temporal characteristic information of PM_(2.5)signals from multiple stations.K-nearest neighbor(KNN)algorithm interpolates the lost signals in the process of collection,transmission,and storage to ensure the continuity of data.Graph generative network(GGN)is used to process time-series meteorological data with complex structures.The graph U-Nets framework is introduced into the GGN model to enhance its controllability to the graph generation process,which is beneficial to improve the efficiency and robustness of the model.In addition,sparse Bayesian regression is incorporated to improve the dimensional disaster defect of traditional kernel density estimation(KDE)interval prediction.With the support of sparse strategy,sparse Bayesian regression kernel density estimation(SBR-KDE)is very efficient in processing high-dimensional large-scale data.The PM_(2.5)data of spring,summer,autumn,and winter from 34 air quality monitoring sites in Beijing verified the accuracy,generalization,and superiority of the proposed model in interval prediction.
基金supported by the National Natural Science Foundation of China(72101263).
文摘Tracking and analyzing data from research projects is critical for understanding research trends and supporting the development of science and technology strategies.However,the data from these projects is often complex and inadequate,making it challenging for researchers to conduct in-depth data mining to improve policies or management.To address this problem,this paper adopts a top-down approach to construct a knowledge graph(KG)for research projects.Firstly,we construct an integrated ontology by referring to the metamodel of various architectures,which is called the meta-model integration conceptual reference model.Subsequently,we use the dependency parsing method to extract knowledge from unstructured textual data and use the entity alignment method based on weakly supervised learning to classify the extracted entities,completing the construction of the KG for the research projects.In addition,a knowledge inference model based on representation learning is employed to achieve knowledge completion and improve the KG.Finally,experiments are conducted on the KG for research projects and the results demonstrate the effectiveness of the proposed method in enriching incomplete data within the KG.
基金supported by the National Key Research and Development Program of China(2023YFA1010200,2020YFA0713100)the National Natural Science Foundation of China(12071453)the Innovation Program for Quantum Science and Technology(2021ZD0302902).
文摘Given two graphs G and H,the Ramsey number R(G,H)is the smallest positive integer N such that every 2-coloring of the edges of K_(N)contains either a red G or a blue H.Let K_(N-1)■K_(1,k)be the graph obtained from K_(N-1)by adding anew vertexνconnecting k vertices of K_(N-1).A graph G withχ(G)=k+1 is called edge-critical if G contains an edge e such thatχ(G-e)=k.A considerable amount of research has been conducted by previous scholars on Ramsey numbers ofgraphs.In this study,we show that for an edge-critical graph G with x(G)=k+1,when k≥2,1≥2,and n is sufficiently large,R(G,K_(1)+nK_(t))=knt+1 and r,(G,K_(1)+nK_(t))=(k-1)nt+1.
文摘A nowhere-zero k-flow on a graph G=(V(G),E(G))is a pair(D,f),where D is an orientation on E(G)and f:E(G)→{±1,±2,,±(k-1)}is a function such that the total outflow equals to the total inflow at each vertex.This concept was introduced by Tutte as an extension of face colorings,and Tutte in 1954 conjectured that every bridgeless graph admits a nowhere-zero 5-flow,known as the 5-Flow Conjecture.This conjecture is verified for some graph classes and remains unresolved as of today.In this paper,we show that every bridgeless graph of Euler genus at most 20 admits a nowhere-zero 5-flow,which improves several known results.
基金supported by the Ministry of Industry and Information Technology(No.23100002022102001)。
文摘Urban combat environments pose complex and variable challenges for UAV path planning due to multidimensional factors,such as static and dynamic obstructions as well as risks of exposure to enemy detection,which threaten flight safety and mission success.Traditional path planning methods typically depend solely on the distribution of static obstacles to generate collision-free paths,without accounting for constraints imposed by enemy detection and strike capabilities.Such a simplified approach can yield safety-compromising routes in highly complex urban airspace.To address these limitations,this study proposes a multi-parameter path planning method based on reachable airspace visibility graphs,which integrates UAV performance constraints,environmental limitations,and exposure risks.An innovative heuristic algorithm is developed to balance operational safety and efficiency by both exposure risks and path length.In the case study set in a typical mixed-use urban area,analysis of airspace visibility graphs reveals significant variations in exposure risk at different regions and altitudes due to building encroachments.Path optimization results indicate that the method can effectively generate covert and efficient flight paths by dynamically adjusting the exposure index,which represents the likelihood of enemy detection,and the path length,which corresponds to mission execution time.
基金supported by the National Natural Science Foundation of China(62176214).
文摘For target tracking and localization in bearing-only sensor network,it is an essential and significant challenge to solve the problem of plug-and-play expansion while stably enhancing the accuracy of state estimation.This paper pro-poses a distributed state estimation method based on two-layer factor graph.Firstly,the measurement model of the bearing-only sensor network is constructed,and by investigating the observ-ability and the Cramer-Rao lower bound of the system model,the preconditions are analyzed.Subsequently,the location fac-tor graph and cubature information filtering algorithm of sensor node pairs are proposed for localized estimation.Building upon this foundation,the mechanism for propagating confidence mes-sages within the fusion factor graph is designed,and is extended to the entire sensor network to achieve global state estimation.Finally,groups of simulation experiments are con-ducted to compare and analyze the results,which verifies the rationality,effectiveness,and superiority of the proposed method.
