Visual Query Language on Spatial Information (SIVQL) is one kind of visual query language based on the extension of Query by Example (QBE). It is a visual operation based on graphics or media object, such as point, li...Visual Query Language on Spatial Information (SIVQL) is one kind of visual query language based on the extension of Query by Example (QBE). It is a visual operation based on graphics or media object, such as point, line and area elements. In this paper, the relation calculation and query function of SIVQL have been studied and discussed by using set theory and relation algebra. The theory foundation of SIVQL has been investigated by the mathematical method. Finally, its application examples are also given with the specific information system.展开更多
大语言模型(large language model,LLM)技术热潮对数据质量的要求提升到了一个新的高度.在现实场景中,数据通常来源不同且高度相关.但由于数据隐私安全问题,跨域异质数据往往不允许集中共享,难以被LLM高效利用.鉴于此,提出了一种LLM和...大语言模型(large language model,LLM)技术热潮对数据质量的要求提升到了一个新的高度.在现实场景中,数据通常来源不同且高度相关.但由于数据隐私安全问题,跨域异质数据往往不允许集中共享,难以被LLM高效利用.鉴于此,提出了一种LLM和知识图谱(knowledge graph,KG)协同的跨域异质数据查询框架,在LLM+KG的范式下给出跨域异质数据查询的一个治理方案.为确保LLM能够适应多场景中的跨域异质数据,首先采用适配器对跨域异质数据进行融合,并构建相应的知识图谱.为提高查询效率,引入线性知识图,并提出同源知识图抽取算法HKGE来实现知识图谱的重构,可显著提高查询性能,确保跨域异质数据治理的高效性.进而,为保证多域数据查询的高可信度,提出可信候选子图匹配算法Trust HKGM,用于检验跨域同源数据的置信度计算和可信候选子图匹配,剔除低质量节点.最后,提出基于线性知识图提示的多域数据查询算法MKLGP,实现LLM+KG范式下的高效可信跨域查询.该方法在多个真实数据集上进行了广泛实验,验证了所提方法的有效性和高效性.展开更多
文摘Visual Query Language on Spatial Information (SIVQL) is one kind of visual query language based on the extension of Query by Example (QBE). It is a visual operation based on graphics or media object, such as point, line and area elements. In this paper, the relation calculation and query function of SIVQL have been studied and discussed by using set theory and relation algebra. The theory foundation of SIVQL has been investigated by the mathematical method. Finally, its application examples are also given with the specific information system.
文摘大语言模型(large language model,LLM)技术热潮对数据质量的要求提升到了一个新的高度.在现实场景中,数据通常来源不同且高度相关.但由于数据隐私安全问题,跨域异质数据往往不允许集中共享,难以被LLM高效利用.鉴于此,提出了一种LLM和知识图谱(knowledge graph,KG)协同的跨域异质数据查询框架,在LLM+KG的范式下给出跨域异质数据查询的一个治理方案.为确保LLM能够适应多场景中的跨域异质数据,首先采用适配器对跨域异质数据进行融合,并构建相应的知识图谱.为提高查询效率,引入线性知识图,并提出同源知识图抽取算法HKGE来实现知识图谱的重构,可显著提高查询性能,确保跨域异质数据治理的高效性.进而,为保证多域数据查询的高可信度,提出可信候选子图匹配算法Trust HKGM,用于检验跨域同源数据的置信度计算和可信候选子图匹配,剔除低质量节点.最后,提出基于线性知识图提示的多域数据查询算法MKLGP,实现LLM+KG范式下的高效可信跨域查询.该方法在多个真实数据集上进行了广泛实验,验证了所提方法的有效性和高效性.
