ontolog-forum
[Top] [All Lists]

Re: [ontolog-forum] IBM Watson on Jeopardy

To: ontolog-forum@xxxxxxxxxxxxxxxx
From: "John F. Sowa" <sowa@xxxxxxxxxxx>
Date: Thu, 10 Feb 2011 15:15:14 -0500
Message-id: <4D544752.1080700@xxxxxxxxxxx>
On 2/10/2011 11:52 AM, Ed Barkmeyer wrote:
> There, John, I think you miss one of the ideas that IBM got right a long
> time ago -- you have to size the computational power to the data flow
> power.    (01)

Ed, I worked at IBM for 30 years.  During the 1980s, IBM had the largest
group of researchers in NLP in the world.  During the cutbacks in the
early '90s, most of them (including me) went elsewhere.  I had also
worked on hardware-software architecture, and I'm very sensitive to
the tradeoffs between the two.    (02)

JFS
>> IBM is doing that with a brute force supercomputer.  But with better
>> algorithms, it's possible to get comparable performance with much
>> less hardware.    (03)

EB
> 'Comparable performance' in a narrow field with a much smaller
> knowledge repository may be possible with much less hardware, but other
> AI technologies can also be very successful in such cases.  Comparable
> performance in a  wider-ranging field with a lot of
> not-clearly-interrelated information will require a large knowledge
> repository and the power to process it.  You are not going to process a
> terabyte knowledge base, or even a 20GB knowledge base, in 60 seconds
> with your twin 4GHz desktop processors.    (04)

Actually, we [VivoMind] are doing that today.  We process terabytes
of documents, translate them to conceptual graphs, compress the graphs
to what we call a Cognitive Signature (TM), index those signatures,
and use logarithmic algorithms to retrieve all those that match
within a given semantic distance.    (05)

The time to create and index the cognitive signatures takes (N log N)
time.  But the retrieval time from a megabyte (10^6) to a terabyte
(10^12) just increases by a factor of 2.    (06)

We can't afford a supercomputer, but we get good performance
on a server with just 8 CPUs.  But the algorithms scale very well,
and some of our clients can afford much bigger machines.    (07)

John    (08)

_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (09)

<Prev in Thread] Current Thread [Next in Thread>