Home » Talks and Events » 2007 » CambridgeSemanticWebGatherings
The Semantic Web provides a common framework that allows data to be shared and reused across application, enterprise, and community boundaries. It is a collaborative effort led by W3C with participation from a large number of researchers and industrial partners.
February 13, 2007 - Cambridge - Massachusetts - USA 
This has been the second chat about the SemanticWeb at MIT in Cambridge. The idea is to get together people in the local area who are interested in and/or working in the Semantic Web. A typical form will be a few presentations of recent work or recent problems discovered, some moderated discussion in the round, and a large amount of unmoderated chat.
At the meeting two demos have been presented :
The first demo was about combining natural language processing with syntactic/semantic graphs. It is possible to ask a question to the system, such question will be translated into graph fragments, such graphs will be checked against a pre-calculated repository of graph fragments related to journal articles. A set of results will be displayed according in order of similarity.
Thus, if I query with the question "What are the main risk factors for dialysis patients with diabetic nephropathy?" I will obtain a graph representing the query:
The results with the related graph fragment are represented in the following pictures. The fragments are matching some sub-parts (at different granularity) of the query graph.
It is interesting to notice what is happening if the query contains some mistakes. For a couple of mistakes like "fr" instaed of "for" and "nephrophaty" instaed of "nephropahy" we will obtain the query graph as follow:
As you can see the graph is changed a bit, but, interestingly the first result is the same as before (figure 2) but the second one, which has the same graph fragment of figure 3 reports only one result instaed of the two of figure 3. The article by the title "Cardiovascular risk in dialysis patients: a comparison of risk factors and cardioprotective therapy between 1996 and 2001" is not even in the foloowing results.
