Semantics and knowledge are keys to realize interconnection environments (Grids). Many researchers are making effort to realize large-scale dynamic networking environment involving both physical network infrastructures and human beings. A basic issue is to support efficient representation, management and sharing of various resources in large-scale environment. It requires fundamental research on the models to understand the characteristics and the evolving principles of the semantic and knowledge networking environment. Scalable platform solutions in the future networking information system should be envisioned. The International Conference on Semantics, Knowledge and Grids (SKG) has been focusing on these topics for five years.
The fifth SKG conference (SKG2009, http://www.knowledgegrid.net/SKG2009) successfully drew researchers from different areas to discuss and explore the possible technical breakthroughs towards the future interconnection environment. Nine representative papers are selected to demonstrate the latest progress in models, methods, platforms and visions on future interconnection environments.
Heterogeneous resource management on large-scale networking environment is always a challenge. Nakanishi1 et al. [1] proposed a knowledge presentation framework for users to browse related heterogeneous knowledge information on the Knowledge Grid [2]. A semantic network method is proposed to represent the related resources in the knowledge base. A dynamic browser is implemented to help users browse knowledge and their related information in a more user-friendly and direct way.
Comito and Talia [3] studied a schema-mapping method of XML data on the P2P network where highly dynamic nodes publish heterogeneous data sources for sharing and integration. A path-to-path schema-mapping method is proposed to reformulate queries and distribute queries among peers to form the query results. The proposed framework can be viewed as a feasible solution for a typical scenario on such large-scale networking environments as the Web.
Zhou et al. [4] proposed an assessment method for evaluating the adaptability of services on the heterogeneous environments where adaptors may not be available. The proposed method is based on graph search and can evaluate whether a requestor and a provider of a service can be matched without using intermediate adaptor. The degree of matching can be measured and the condition of matching can be identified when the matching is possible.
Automatically discovering and mapping ontologies on the network is a key problem for building intelligent Knowledge Grid. Mao et al.
[5] used a machine-learning technique to address the ontology-mapping problem. The ontology-mapping problem is mapped onto a binary classification problem that many traditional machine-learning methods can apply. To make the classification more generic, the authors selected a set of features that are independent to specific instances and domains. The features can be applied to different domains without using instance training. This advantage can be promising for implementing semantic search/retrieval on large-scale Semantic Web data set.
Classification and link are two basic means that connect various resources in future interconnection environment [6]. Discovering semantic links is important for intelligently sharing information. Zhuge and Zhang [7] proposed a rule-based semantic link discovery method on a semantic link network formed by document contents. The proposed method supports many important features including probabilistic relational reasoning, semantic link network evolution and rule evolution. It can be applied to the semantics mining of text contents. It can be also applied to other networked semantic data such as the Semantic Web since the method incorporates rules to guide the semantic link discovery.