Semantic Web a reality

If you're new here, you may want to subscribe to my RSS feed. Thanks for visiting!

Slashdot reports TimO’Reilly interview with Devin Wning, CEO, Reuters. It talks about the fast and smart news which not only offers the news without any latency but also does it smartly by automatically detecting linkages in it ; name, place and things. Reuters has provided free and open access to Calais API, which turns returns a formal meaningful RDF graph from an unstructured text.

Softlab adds:

Secondly, Wenig claimed that we are coming to the end of an era where the company with the least time delay in delivering news held a competitive advantage. This second point exposed a very important trend for the future of news data: that the timing of news is no longer a crucial factor, but rather the sources of the news and the information which can be derived from connections between them. In other words, the processing of the data. This is where the Semantic Web steps in. The aim is not just to mark data with semantic metadata, but to use the semantic data to derive added-value additional information from the original data for the consumer, where the consumer may be another news company, or the end consumer. Thus, the focus is on making insights from the data through semantic technology.

ReadWriteWeb has some more information on Calais API and how it matters for Reuters.

The future of news thus in not limited providing or publishing the news but to make the connection and provide much more insight into the news. And as the news is not just limited to a textual story, the outcome out of a semantic news will be pretty complex, challenging but very very interesting. On a semantic news platform the thing has to make sense through the main participants of the news - name, place, things (date, people, location, time, company, etc) and as the also through the various media formats available through the news (audio, video, pictures and text).

Generating semantics through text itself is not easy and making sense of media would add to the overall challenge. Would this be done by the publisher of the media or the extraction through this can be done machines ?. Definitely it means lots and lots of processing and has to start with seeding of metadata from the publishers and using the to build up the AI through natural language processing and learning.

Popularity: 1% [?]