Ontologies

Ontology is the backbone of Semantic Web technology. How to generate ontologies semi-automatically and to create on-site mapping and versioning of ontologies is a critical line of research.

  • Ontology Engineering: This line of research focuses on manual-creation of ontologies by applying various knowledge acquisition methods (e.g., interviewing, self-reporting, laddering, concept sorting, repertory grids, automatic learning techniques, etc.), knowledge modeling technologies (e.g., modularization, top-level ontologies, spiral knowledge model, etc.) and existing ontology engineering methods (e.g., TOVE, Methotology, etc.)
  • Ontology Generation: This research investigates the semi-automatic use of linguistic support to extract a part of a domain or application ontology semi-automatically (e.g., classes of ontologies normalized as key noun phrases). Relation extraction is one of the primary bottlenecks in ontology generation, and associations among noun phrases have been identified as one way to extract relationships within these phrases. Future research will focus on extraction of relationships among verbs, adjectives and/or prepositions for automatic generation of ontologies.
  • Ontology mediation: Ontology meditation is key part alignment of ontologies and the management of ontology structures. Current solutions for ontology mediation remain at the stage of manually aligning and mapping ontologies, with some limited recommendation services. Ontology mediation research is currently focused on the identification of patterns for ontology mediation, the development of ontology mediation libraries, and the storage of identified patterns in mediation libraries that will support the grouping of patterns in "pattern clusters" and allow for flexible and easy access to and reuse of these patterns. Additionally, personalized views of mediation patterns could be tailored according to the requirements of the specific task or application.

Social Network Analysis for Research Impact

Social network analysis considers the topology of networks when ranking network nodes. It offers another important approach for measuring the impact of scholars in scholarly communications. This research focuses on how to use social network analysis approaches to evaluate research impact, including various centrality measures, network features, PageRank, and weighted PageRank. The goal of this line of research is to add temporal features to current PageRank algorithms and to identify the potential convergence of impact measures.

Social Tagging

Tagging has become more and more popular for web users to organize and share the data and resources they have stored on the Web. It offers an interesting possibility for social indexing and the representation of Web resources. This line of research focuses on how to improve searching with tags (e.g., cross-platform, cross-network) based on tag ontologies and how to mine associations among tagging data. Potential applications are identifying personal profiles, use in recommender systems, and tagging or annotating in digital libraries, etc. In the future, TagULike -- an application similar to CiteULike or CiteSeer that would replace citation data with tagging information -- might extend citation data on the web and become a potential space for interesting research and exploitation.

Semantic Web for Biomedicine

Linked Open Data (LOD) provides linked semantic data for the public. Bio2RDF converts most important Biodata into RDF. It opens another research opportunity that investigates how to integrate Bio2RDF data with experimental data and workflows in order to trace provenance, semantically mine the bio RDF graphs, and provide provenance aware visualizations.

Data integration and mediation in Web2.0

This work investigates adding metadata to the current Web through tagging so that different blogs, wikis and BBS can be integrated and communicated. This work is based on the integration of different existing popular metadata and/or social ontologies, such as FOAF (Friend of A Friend), DC (Dublin Core Metadata), SKOS (Simple Knowledge Organization Systems), SIOC (Semantically-Interlinked Online Communities), RSS, etc., to hyperlink resources for searching and mining.

Scientific Profiling and Evaluating

This line of research involves the use of large amounts of crawled web data to profile and evaluate scientific performance of certain organizations based on scientometrics and bibliometrics and is undertaken in collaboration with The Library of Chinese Academy of Sciences.