Ontology Population and Concept Search across Ontologies: Challenges and Methodologies

Semantic Web has brought a new paradigm to web search. The primary components of the Semantic Web, ontologies and knowledge graphs (populated ontologies), are rich
sources of domain knowledge. Semantic search systems leverage this domain knowledge to capture query intent and improve the accuracy of the search. The research addresses the key aspects of semantic search: (1) creation of the knowledge graphs and (2) efficient retrieval of concepts across distributed and interlinked ontologies.

We propose efficient reuse of structured and unstructured resources to build knowledge graphs. The manual method of populating ontologies with instances to build knowledge graphs is laborious. We leverage bootstrapping techniques to reuse unstructured resources for semiautomatic ontology population. In addition, instances were extracted from the structured resources on the linked open data. Our semiautomatic ontology population methodologies have the potential to reduce the time and effort required to build the knowledge graph.

The second key aspect of semantic search is retrieving concepts from a corpus of ontologies. Identifying the right concepts is a critical part of search-improvement techniques such as web page annotation, query intent capture, and web page classification. Concept search is also useful in other real-world applications such as ontology mapping, ontology merging and link prediction. Collectively, tens of millions of concepts (Berners-Lee et al., 2001; Bauer and Kaltenbock, 2011) have been defined in a large number of ontologies that cover many overlapping domains. The scale, duplication and ambiguity makes concept search across ontologies a challenging problem. We propose a keyword-based concept search framework that balances the relevance and diversity of concept search results.