XLike
Aus Aifbportal
 |
Cross Lingual Knowledge Extraction |
|
|
The goal of the X-LIKE project is to develop technology to monitor and aggregate knowledge that is currently spread across global mainstream and social media, and to enable cross-lingual services for publishers, media monitoring and business intelligence. In terms of research contributions, the aim is to combine scientific insights from several scientific areas to contribute in the area of cross-lingual text understanding. By combining modern computational linguistics, machine learning, text mining and semantic technologies we plan to deal with the following two key open research problems: - to extract and integrate formal knowledge from multilingual texts with cross-lingual knowledge bases, and - to adapt linguistic techniques and crowdsourcing to deal with irregularities in informal language used primarily in social media. As an interlingua, knowledge resources from Linked Open Data cloud (http://linkeddata.org/) will be used with special focus on general common sense knowledge base CycKB (http://www.cyc.com/). For the languages where no required linguistic resources will be available, we will use a probabilistic interlingua representation trained from a comparable corpus drawn from the Wikipedia. The solution will be applied on two case studies, both from the area of news. For the Bloomberg case study the domain will be financial news, while for the Slovenian Press Agency we will deal with general news. The technology developed in the project will be used to introduce cross-lingual and information from social media in services for publishers and end-users in the area of summarization, contextualization, personalization, and plagiarism detection. Special attention will be paid to analysing news reporting bias from multilingual sources. The developed technology will be language-agnostic, while within the project we will specifically address English, German, Spanish, and Chinese as major world languages and Catalan and Slovenian as minority languages.
von: 1 Januar 2011
bis: 31 Dezember 2014
Finanzierung: EU
INSTITUT JOZEF STEFAN, UNIVERSITAT POLITECNICA DE CATALUNYA, SVEUCILISTE U ZAGREBU, TSINGHUA UNIVERSITY, INTELLIGENT SOFTWARE COMPONENTS S.A., SLOVENSKA TISKOVNA AGENCIJA DOO
Text Mining, Natürliche Sprachverarbeitung, Maschinelles Lernen, Semantische Technologien
inproceedings
Uta Lösch, Stephan Bloehdorn, Achim Rettinger
Graph Kernels for RDF Data
In Elena Simperl et. al., Proc. of the 9th Extended Semantic Web Conference (ESWC'12), Springer, Mai, 2012
(Details)
↑ top
