Looking at the enormous growth of available knowledge and accessable data, methods for structuring and extracting information are more and more important as a future field of value generation.
The Living Knowledge project aims at bringing a new quality into search and knowledge management technology for more concise, complete and contextualised results. - More than 6.000 registered users at the project website demonstrate considerable interest in the topic and the ongoing development of test applications.
SORA provides the social-science exertise in the framework of the interdisciplinary project consortium.
Application Scenario: Future Predictor
The future predictor application searches for future predictions (plans or estimates) in current documents, tracks them over time and supports reasoning on the impact of such future predictions.
By integrating the outputs of the Living Knowledge partners with an intuitive user interface, the application allows users to learn about what people are saying about the future. What are the hot topics? Is the sentiment positive or negative? What are the information sources and are they biased in anyway? Crucially, the user is able to explore how all of these aspects evolve over time and space. An early prototype of the system was selected as the People’s Choice award at the 2010 HCIR Challenge and featured in the MIT Technology Review.
The system was developed as part of the Living Knowledge project by the Natural Language Retrieval Group at Yahoo! Research Barcelona.
T-YAGO: Boosting intelligence in knowledge data bases
In order to effectively exploit the enormous knowledge available in high-quality Internet sources such as Wikipedia, one needs to extract and structure information and develop efficient methods for querying and exploration. YAGO is a large-scale and highly accurate knowledge base with convenient search capabilities.
The Living Knowledge team harvested facts about individual entities from infoboxes and categories of Wikipedia articles and unified them with WordNet. Currently YAGO contains about 3 million entities and more than 20 million facts (instances of binary relations between entities).
Building upon this database, T-YAGO for the first time adds a time index to information and makes it possible to display knowledge and its evoluation on a time axis (this T-YAGE or "temporal" YAGO).
Potential applications range from the temporal monitoring of print media or entries in the Web 2.0 to depicting trends and relations among all kinds of entities. Queries in T-YAGO could yield information on eduction, positions occupied over time, but also relations to other people, firms, locations or institutions - as well as their evolution over time.
More information: livingknowledge-project.eu