torek, 1. junij 2010

Wolfram Alfa: Computing a theory of everything

I came across a lecture from Stephen Wolfram about a computational knowledge engine, called Wolfram Alfa. It is his lifelong project, that is still not quite finished, but still, I find his idea and the project very interesting.
Stephen Wolfram wanted to see how much of the systematizing knowledge that is ''out there in the world'' and somehow make it computable. So last year, after much hard work, they released the first website version of Wolfram Alpha. This is a serious knowledge engine that computes answers to questions. You type in what you want to know (for example Slovenian population or, a little bit more difficult, the difference between Dutch and Slovenian population). It covers all the areas you can think of: health, geography, economy, math etc. Steven Wolfram says it covers everything one might find in a standard reference library, but ''its goal is to democratize all of this kind of knowledge, and try to bea n authoritative source in all areas, to be able to compute answers to specific questions that people have, not by searching what other people may have written down before, but by using built in knowledge to compute fresh new answers to specific questions.'' He continues: ‘’Wolfram Alpha actually gives one a sort of whole new kind of computing that one can call knowledge-based computing, in which one's starting, not just from raw computation, but from a vast amount of built-in knowledge. And when one does that, one really changes the economics of delivering computational things, whether it's on the web or elsewhere.’’
You can find his whole, 20 minute long lecture on this website: http://www.ted.com/talks/stephen_wolfram_computing_a_theory_of_everything.html.

I find his idea very attractive, but I am somewhat skeptical on this point. The fact is that Wolfram Alpha is not a ordinary search engine that you can find on the Internet. It is a computational knowledge engine, which means that it generates output by doing computations from its own internal knowledge base, instead of searching the web and returning links. So it only knows what is in those bases and has to rely on people to input all the pieces of knowledge in a sufficiently form in order to make processing possible. But all-in-all, the idea of a computational knowledge engine is really great and in a couple of years, when it is fully developed, it will make a great contribution to the Internet and to the society as a whole.

Ni komentarjev:

Objavite komentar