- Spela Vintar, DFKI, Saarbrücken, zum Thema Semantic Processing for Cross-Language Information Retrieval. Ort: Komplex Golm, Haus 24, R.075. Zeit: Dienstag, 18.Juni '02, 17 Uhr s.t.
Abstract
In a time of advancing Semantic Web technologies, using conceptual data on top of
standard linguistic processing has become an area of interest not only within the
NLP community but also in the broader context of knowledge access and engineering.An
ontology is the basic resource for such applications, and developing advanced
methods of using and expanding ontologies for Document Retrieval is the research
goal of several projects and groups in various domains.
The talk will briefly introduce the methods used in concept-based IR/CLIR and then
describe the experience gained so far in the MUCHMORE (Multilingual Concept
Hierarchies for Medical Information Organisation and Retrieval) project. We use
several levels of semantic annotation based on the UMLS (Unified Medical Language
System) and explore ways of tuning and expanding semantic processing by integrating
Word Sense Disambiguation, Term Extraction and Extraction of Semantic Relations. The
talk will briefly present the tools and methods used for the above tasks and show
examples of annotated medical texts. Some unresolved problems of semantic processing
will also be discussed.