« Return to Thread: Wolfram Alpha

Re: Wolfram Alpha

by John F. Sowa :: Rate this Message:

Reply to Author | View in Thread

Stephen Wolfram presented a "sneak preview" of the Alpha system
at Harvard on Tuesday.  The You Tube version has been posted on
the Harvard web site:

    http://cyber.law.harvard.edu/events/2009/04/wolfram

For the first half hour, Stephen W. gave a demo of Alpha in action.
Like most demos, 5 minutes was good, 10 minutes was too long, and
30 minutes was boring.  But then he switched to Q/A mode, which
was much more interesting.

As the examples and the discussion indicated, the Alpha system
has a large number of templates for interpreting typical question
patterns by users.  Those templates are linked to Mathematica
_notebooks_, which correspond to Cyc microtheories, but with
everything represented in Mathematica notation and formalisms.

In the Q/A session, somebody asked for a comparison with Cyc, and
Wolfram said that Alpha is complementary.  Unlike Cyc, which was
designed for shallow commonsense reasoning about a broad range
of topics, Alpha is designed to do very deep computation about
many very narrow areas -- chemistry, medicine, geography,
finance, etc.

In effect, Alpha has a lot of special case ontologies specified
in the notebooks.  The only kind of common information is about
mathematics and units of measures.  The selection of a specific
notebook is done by the language templates and words.  Wolfram
noted that words can often be misleading, as in the case of a
person named 50 cents.

They plan to make the Alpha system available on the WWW later this
month.  There is an enormous difference between a demo with the
developer at the keyboard, and a blank screen where users can type
anything that comes to mind.  The test by actual users will be the
critical one.

John Sowa


---------------------------------------------------------------------
To unsubscribe, e-mail: cg-unsubscribe@...
For additional commands, e-mail: cg-help@...

 « Return to Thread: Wolfram Alpha