OntoSum - Ontology Management and Use to Support Summarization

Printer-friendly versionSend by email
Ontology Management and Use to Support Summarization

The OntoSum project is funded by the Greek government and supports three PhD students working on the following subjects:

  • Ontology Learning from textual content and data bases: This concerns the development of methods that enable the enrichment of an initial ontology (e.g. the discovery of new concepts, new properties or relationships of existing concepts, update of existing concepts, properties, relationships). Such methods involve techniques from the areas of machine learning, natural language processing, knowledge representation and management.
  • Ontology Coordination: This concerns the cases where more than one ontologies for the same domain must be considered. These ontologies may represent different types of data or they may represent the same types of data but containing different concepts, properties, relationships, etc. (different not only in names used but also in what they really represent in the context of the specific domain since they may have been constructed by different domain experts). Ontology coordination methods are used to handle such cases. In an ontology coordination process, mapping, alignment and merging are the major sub-processes. Coordination can be performed using either a reference ontology or directly, but maintaining the semantic of each ontology at the end.
  • Ontology-based Multi Document Summarization: This concerns the development of automatic methods for summarising documents (textual documents and/or data from data bases) talking abouta specific event. For instance, corporate news for management succession events (people leaving companies, changing positions, retiting, etc.) in a specific period in time. Summarisation methods may involve techniques for document classification (to locate the documents discussing the specific events), information extraction from the retrieved documents (in order to extract the main features for the specific events), language generation (the extracted data may be store din a database and you use natural language generation techniques to generate the summary from those data). In all the above processing steps, the domain specific ontology has a crucial role driving the classification, extraction and generation tasks.
GSRT/Ministry of Development
Project type: 
Start to End date: 
01/12/2005 - 30/11/2008
43 months
Financial Info
Project contract: 
Funding organization: 
GEL code: 
Total budget: 
NCSR budget: 
EU funding: 
Matching fund: 

© 2018 - Institute of Informatics and Telecommunications | National Centre for Scientific Research "Demokritos"