Gastvortrag "From Information Extraction to Semantic Search"

Moderator: Semantic Web

Beiträge: 89
Registriert: 4. Jun 2008 09:07

Gastvortrag "From Information Extraction to Semantic Search"

Beitrag von HeikoPaulheim »

Hallo zusammen,

morgen nach der Vorlesung, von 11:30-13:00 in A126, hält Dr. Roman Klinger vom Fraunhofer Institute for Algorithms and Scientific Computing einen Gastvortrag zum Thema "From Information Extraction to Semantic Search".

Vielleicht hat ja jemand Interesse.

Viele Grüße,
Heiko Paulheim.

Textual databases of scientific publications are growing fast such that for researchers it can be hard to keep track with the development in a specific field. As more and more publications become available for data and text mining purposes, information extraction can support the task of structuring the knowledge presented in natural language. I briefly introduce the state-of-the-art machine learning method of conditional random fields and present a use case from the domain of bio-chemistry. Specific challenges to apply a trained model to large corpora are discussed, especially feature selection approaches. Our strategy to analyze huge corpora in a high-performance computing setting and making the result available as a semantic search engine based on Lucene are presented.

Zurück zu „Semantic Web“