Semantic Web gets underway
I have been a beta tester for True Knowledge, a UK company which aims to build a repository of knowledge contributed by users and other sources. Before you dismiss them as another Wikipedia clone, True Knowledge is built on semantic web concepts such as the RDF triple of subject, predicate and object.
This means that apart from adding new facts to the repository, users are required to provide True Knowledge with relationships between objects that make up that fact which other users can endorse or refute. Unlike hyperlinked knowledge systems like Wikipedia, True Knowledge machines are able to process and understand knowledge as we know it.
One of the first things I did was to include Singapore in the geographical region of Southeast Asia. The system prompted me about the sources for this information and I provided a link to an atlas. I also endorsed a fact which someone else entered stating the population of Singapore as 4490000, as of 2006. Later on, I added that Singapore has been a republic since August 9, 1965.
It's interesting how the system is able to give you the necessary prompts to confirm relationships and the definitions of objects among relationships so you know exactly what you're saying. As a result of this, True Knowledge is able to understand your questions in natural language as opposed to matching a string of text against an inverted index. Here's the results when I queried for "Where is Singapore?"
Apart from user contributed knowledge, True Knowledge also mines information from various sources such as Wikipedia and the CIA World Fact Book. If you search for Lee Kuan Yew, Singapore's first Prime Minister, you get his details found in a Wikipedia profile.
Powered by knowledge representation systems that melds the wisdom of crowds with authoritative sources, semantic searches like these make our library systems look truly archaic.
0 comments:
Post a Comment