Architecture framework has become an effective method recently to describe the system of systems(SoS)architecture,such as the United States(US)Department of Defense Architecture Framework Version 2.0(DoDAF2.0).As a vi...Architecture framework has become an effective method recently to describe the system of systems(SoS)architecture,such as the United States(US)Department of Defense Architecture Framework Version 2.0(DoDAF2.0).As a viewpoint in DoDAF2.0,the operational viewpoint(OV)describes operational activities,nodes,and resource flows.The OV models are important for SoS architecture development.However,as the SoS complexity increases,constructing OV models with traditional methods exposes shortcomings,such as inefficient data collection and low modeling standards.Therefore,we propose an intelligent modeling method for five OV models,including operational resource flow OV-2,organizational relationships OV-4,operational activity hierarchy OV-5a,operational activities model OV-5b,and operational activity sequences OV-6c.The main idea of the method is to extract OV architecture data from text and generate interoperable OV models.First,we construct the OV meta model based on the DoDAF2.0 meta model(DM2).Second,OV architecture named entities is recognized from text based on the bidirectional long short-term memory and conditional random field(BiLSTM-CRF)model.And OV architecture relationships are collected with relationship extraction rules.Finally,we define the generation rules for OV models and develop an OV modeling tool.We use unmanned surface vehicles(USV)swarm target defense SoS architecture as a case to verify the feasibility and effectiveness of the intelligent modeling method.展开更多
针对空间有效载荷系统高复杂性和高可靠性需求的特性,设计了一种基于SysML (System Modeling Language)的故障诊断方法.该方法融入MBSE (Model Based System Engineering)思想,提出了基于SysML的空间有效载荷系统故障分析流程.基于SysM...针对空间有效载荷系统高复杂性和高可靠性需求的特性,设计了一种基于SysML (System Modeling Language)的故障诊断方法.该方法融入MBSE (Model Based System Engineering)思想,提出了基于SysML的空间有效载荷系统故障分析流程.基于SysML对空间有效载荷系统建立了故障分析相关的模型,其中,为满足故障分析建模的需求,对SysML元模型进行扩展定义,从而实现对组件间关系和故障表征与直接关联组件间关系的描述;基于所建模型构建故障诊断的整体框架,并提供从SysML数字模型到FTA (Fault Tree Analysis)的转换逻辑,从而实现对所有故障可能性的获取.通过案例分析,对提出方法在实际应用中的具体流程进行分析,并验证了该方法的有效性和实用性.展开更多
The development process of complex equipment involves multi-stage business processes,multi-level product architecture,and multi-disciplinary physical processes.The relationship between its system model and various dis...The development process of complex equipment involves multi-stage business processes,multi-level product architecture,and multi-disciplinary physical processes.The relationship between its system model and various disciplinary models is extremely complicated.In the modeling and integration process,extensive customized development is needed to realize model integration and interoperability in different business scenarios.Meanwhile,the differences in modeling and interaction between different modeling tools make it difficult to support the consistent representation of models in complex scenarios.To improve the efficiency of system modeling and integration in complex business scenarios,a system modeling and integration method was proposed.This method took the Sys ML language kernel as the core and system model function integration as the main line.Through the technical means of model view separation,abstract operation interface,and model view configuration,the model modeling and integration of multi-user,multi-model,multi-view,and different business logic in complex business scenarios were realized.展开更多
大语言模型(LLMs,Large Language Models)具有极强的自然语言理解和复杂问题求解能力,本文基于大语言模型构建了矿物问答系统,以高效地获取矿物知识。该系统首先从互联网资源获取矿物数据,清洗后将矿物数据结构化为矿物文档和问答对;将...大语言模型(LLMs,Large Language Models)具有极强的自然语言理解和复杂问题求解能力,本文基于大语言模型构建了矿物问答系统,以高效地获取矿物知识。该系统首先从互联网资源获取矿物数据,清洗后将矿物数据结构化为矿物文档和问答对;将矿物文档经过格式转换和建立索引后转化为矿物知识库,用于检索增强大语言模型生成,问答对用于微调大语言模型。使用矿物知识库检索增强大语言模型生成时,采用先召回再精排的两级检索模式,以获得更好的大语言模型生成结果。矿物大语言模型微调采用了主流的低秩适配(Low-Rank Adaption,LoRA)方法,以较少的训练参数获得了与全参微调性能相当的效果,节省了计算资源。实验结果表明,基于检索增强生成的大语言模型的矿物问答系统能以较高的准确率快捷地获取矿物知识。展开更多
随着全球气候变化日益严重,企业碳排放分析成为国际关注的焦点,针对通用大语言模型(large language model,LLM)知识更新滞后,增强生成架构在处理复杂问题时缺乏专业性与准确性,以及大模型生成结果中幻觉率高的问题,通过构建专有知识库,...随着全球气候变化日益严重,企业碳排放分析成为国际关注的焦点,针对通用大语言模型(large language model,LLM)知识更新滞后,增强生成架构在处理复杂问题时缺乏专业性与准确性,以及大模型生成结果中幻觉率高的问题,通过构建专有知识库,开发了基于大语言模型的企业碳排放分析与知识问答系统。提出了一种多样化索引模块构建方法,构建高质量的知识与法规检索数据集。针对碳排放报告(政策)领域的知识问答任务,提出了自提示检索增强生成架构,集成意图识别、改进的结构化思维链、混合检索技术、高质量提示工程和Text2SQL系统,支持多维度分析企业可持续性报告,为企业碳排放报告(政策)提供了一种高效、精准的知识问答解决方案。通过多层分块机制、文档索引和幻觉识别功能,确保结果的准确性与可验证性,降低了LLM技术在系统中的幻觉率。通过对比实验,所提算法在各模块的协同下在检索增强生成实验中各指标表现优异,对于企业碳排放报告的关键信息抽取和报告评价,尤其是长文本处理具有明显的优势。展开更多
近年来,以Chat GPT为代表的大语言模型(large language model,LLM)技术发展迅速.随着模型参数规模的持续增长,构建和应用大模型对数据存储规模和存储访问效率提出了更高要求,这对传统存储系统带来了严峻挑战.首先分析了大模型在数据准...近年来,以Chat GPT为代表的大语言模型(large language model,LLM)技术发展迅速.随着模型参数规模的持续增长,构建和应用大模型对数据存储规模和存储访问效率提出了更高要求,这对传统存储系统带来了严峻挑战.首先分析了大模型在数据准备、模型训练和推理阶段的存储访问特征,深入探讨了传统存储系统在大模型场景下面临的主要问题和瓶颈.针对这些挑战,提出并实现了一种高性能、可扩展的分布式元数据设计Scale FS.通过目录树元数据与属性元数据解耦的架构设计,并结合深度与广度均衡的目录树分层分区策略设计,Scale FS实现了高效的路径解析、负载均衡和系统扩展能力,能够高效管理千亿级文件.此外,Scale FS设计了细粒度元数据结构,优化了元数据访问模式,并构建了面向文件语义优化的元数据键值存储底座,显著提升了元数据访问效率并减少了磁盘I/O操作.实验结果表明,Scale FS的每秒操作次数(operations per second,OPS)是HDFS的1.04~7.12倍,而延迟仅为HDFS的12.67%~99.55%.在千亿级文件规模下,Scale FS的大部分操作性能优于HDFS在十亿级文件规模下的表现,展现出更高的扩展性和访问效率,能够更好地满足大模型场景对千亿级文件存储及高效访问的需求.展开更多
基金National Natural Science Foundation of China(71690233,71971213,71901214)。
文摘Architecture framework has become an effective method recently to describe the system of systems(SoS)architecture,such as the United States(US)Department of Defense Architecture Framework Version 2.0(DoDAF2.0).As a viewpoint in DoDAF2.0,the operational viewpoint(OV)describes operational activities,nodes,and resource flows.The OV models are important for SoS architecture development.However,as the SoS complexity increases,constructing OV models with traditional methods exposes shortcomings,such as inefficient data collection and low modeling standards.Therefore,we propose an intelligent modeling method for five OV models,including operational resource flow OV-2,organizational relationships OV-4,operational activity hierarchy OV-5a,operational activities model OV-5b,and operational activity sequences OV-6c.The main idea of the method is to extract OV architecture data from text and generate interoperable OV models.First,we construct the OV meta model based on the DoDAF2.0 meta model(DM2).Second,OV architecture named entities is recognized from text based on the bidirectional long short-term memory and conditional random field(BiLSTM-CRF)model.And OV architecture relationships are collected with relationship extraction rules.Finally,we define the generation rules for OV models and develop an OV modeling tool.We use unmanned surface vehicles(USV)swarm target defense SoS architecture as a case to verify the feasibility and effectiveness of the intelligent modeling method.
文摘针对空间有效载荷系统高复杂性和高可靠性需求的特性,设计了一种基于SysML (System Modeling Language)的故障诊断方法.该方法融入MBSE (Model Based System Engineering)思想,提出了基于SysML的空间有效载荷系统故障分析流程.基于SysML对空间有效载荷系统建立了故障分析相关的模型,其中,为满足故障分析建模的需求,对SysML元模型进行扩展定义,从而实现对组件间关系和故障表征与直接关联组件间关系的描述;基于所建模型构建故障诊断的整体框架,并提供从SysML数字模型到FTA (Fault Tree Analysis)的转换逻辑,从而实现对所有故障可能性的获取.通过案例分析,对提出方法在实际应用中的具体流程进行分析,并验证了该方法的有效性和实用性.
文摘The development process of complex equipment involves multi-stage business processes,multi-level product architecture,and multi-disciplinary physical processes.The relationship between its system model and various disciplinary models is extremely complicated.In the modeling and integration process,extensive customized development is needed to realize model integration and interoperability in different business scenarios.Meanwhile,the differences in modeling and interaction between different modeling tools make it difficult to support the consistent representation of models in complex scenarios.To improve the efficiency of system modeling and integration in complex business scenarios,a system modeling and integration method was proposed.This method took the Sys ML language kernel as the core and system model function integration as the main line.Through the technical means of model view separation,abstract operation interface,and model view configuration,the model modeling and integration of multi-user,multi-model,multi-view,and different business logic in complex business scenarios were realized.
文摘大语言模型(LLMs,Large Language Models)具有极强的自然语言理解和复杂问题求解能力,本文基于大语言模型构建了矿物问答系统,以高效地获取矿物知识。该系统首先从互联网资源获取矿物数据,清洗后将矿物数据结构化为矿物文档和问答对;将矿物文档经过格式转换和建立索引后转化为矿物知识库,用于检索增强大语言模型生成,问答对用于微调大语言模型。使用矿物知识库检索增强大语言模型生成时,采用先召回再精排的两级检索模式,以获得更好的大语言模型生成结果。矿物大语言模型微调采用了主流的低秩适配(Low-Rank Adaption,LoRA)方法,以较少的训练参数获得了与全参微调性能相当的效果,节省了计算资源。实验结果表明,基于检索增强生成的大语言模型的矿物问答系统能以较高的准确率快捷地获取矿物知识。
文摘随着全球气候变化日益严重,企业碳排放分析成为国际关注的焦点,针对通用大语言模型(large language model,LLM)知识更新滞后,增强生成架构在处理复杂问题时缺乏专业性与准确性,以及大模型生成结果中幻觉率高的问题,通过构建专有知识库,开发了基于大语言模型的企业碳排放分析与知识问答系统。提出了一种多样化索引模块构建方法,构建高质量的知识与法规检索数据集。针对碳排放报告(政策)领域的知识问答任务,提出了自提示检索增强生成架构,集成意图识别、改进的结构化思维链、混合检索技术、高质量提示工程和Text2SQL系统,支持多维度分析企业可持续性报告,为企业碳排放报告(政策)提供了一种高效、精准的知识问答解决方案。通过多层分块机制、文档索引和幻觉识别功能,确保结果的准确性与可验证性,降低了LLM技术在系统中的幻觉率。通过对比实验,所提算法在各模块的协同下在检索增强生成实验中各指标表现优异,对于企业碳排放报告的关键信息抽取和报告评价,尤其是长文本处理具有明显的优势。
