中文摘要：

提出一种从虚拟组织自治域的资源描述中抽取语义，然后聚合为虚拟组织的语义基础设施的方法．该方法引入了一种领域知识学习算法，用以建立当前语境相关的词法空间，以提高语义抽取和聚合的准确性及自动化程度，并且在语义聚合的过程中隐含了虚拟组织语义到自治域语义的映射，更好的支持了虚拟组织应用的构建和跨自治域资源的透明访问．实验表明，该方法能够适应虚拟组织的动态开放环境、有效支持虚拟组织的语义基础设施构建．

英文摘要：

The paper proposes an approach to developing the semantic infrastructure for virtual organizations by extracting semantics from resource descriptions in individual autonomous domains and then fusing the extractions. The approach makes the extracted and fused semantics more accurate by using an algorithm to identify the lexical space relative to current scenario and enhances automaticity of the semantic extraction and fusion. The mappings between the semantics of a virtual organization and the corresponding autonomous domains is reflected by the course of semantics fusion, therefore, the approach can well support to integrate applications for virtual organizations and transparently access heterogeneous resources across autonomous domains. Experiments show that the approach suits the dynamic and opening environment of virtual organizations by supporting the development of the semantic infrastructures.