Sunday, May 30, 2010

Super computational engine Wolf|ram Apha - a review

In its FAQ page, developers of Wolfram Alpha are quick to point out that Wolfram Alpha is not a search engine. It is described instead, as a computational knowledge engine. When a user plugs a query into the search box provided, WA generates output by doing computations and providing answers using its built-in knowledge base and algorithms, instead of searching the web and returning links.

As of now, WolframAlpha contains 10+ trillion pieces of data, 50,000+ types of algorithms and models, and linguistic capabilities for 1000+ domains. All data in WolframAlpha is centrally curated and audited bythe WA Team. This team also use a portfolio of automated and manual methods to check data, including statistics, visualization, source cross-checking, and expert review.

The lofty goal of this computational engine is to take as much of the world knowledge as possible and make it computable and accessible to everyone. Taking all the data, methods, models and algorithms accumulated during our civilization and make this immediately computable so that anyone, anywhere, can go on the web and use that knowledge to compute answers to their specific questions.

Their target group is indeed wide-ranging spanning all professions and education levels. Any level, from kindergarten to graduate school to a practitioner in the field can use this engine to get answers. For example for kindergarten students, WolframAlpha can do simple arithmetic what is 2+2, showing a step by step calculation. For the graduate student, what is the GDP of France, for the researcher how many internet users in Europe?

This seemingly super computational engine does list some of its main limitations as follows: engine can only know things that are known, and are somehow public; only deals with facts, not opinions; limits the computation time for each query; and of course, it's in continual development (continuous beta).

Click here to view tutorial video