ontolog-forum
[Top] [All Lists]

Re: [ontolog-forum] Watson going mass market - jobs for ontologists

To: ontolog-forum@xxxxxxxxxxxxxxxx
From: John F Sowa <sowa@xxxxxxxxxxx>
Date: Fri, 22 Nov 2013 12:16:13 -0500
Message-id: <528F915D.6010903@xxxxxxxxxxx>
Ron, Kingsley, Jack, and Simon,    (01)

RW
> Is this going to be the big breakthrough that makes ontology a big deal?    (02)

KI
> Yes!    (03)

I'd be more cautious about the exclamation point, but I'll certainly
say "Can't hurt, might help."    (04)

Note the footnote at the end of the article from the OIL IT Journal:
> UIMA apart, Frase failed to make clear the contribution of open
> source software to Watson. According to Heaton Research
> <http://www.oilit.com/links/1112_0104>, Watson runs on Linux and Hadoop.
> IBM reported a $1 billon/year investment
> <http://www.oilit.com/links/1112_0105> in open source back in 2000
> and produces the excellent IBM Developer Works
> <http://www.oilit.com/links/1112_0106> resource.    (05)

If anyone has any doubts that Watson will be integrated with open source
software, note that Jim Hendler and his students and colleagues at RPI
received a "modified version" of Watson to play with.    (06)

 From http://watson.rpi.edu/
> On January 30, 2013, IBM announced it will provide a modified version
> of an IBM Watson system to Rensselaer, making this the first university
> to receive such a system.    (07)

JP
> This article http://www-03.ibm.com/press/us/en/pressrelease/42451.wss
> refers to:  IBM Watson Developers Cloud    (08)

Yes, and I like the phrase "Fueling New Era of Cognitive Apps".    (09)

Note that Watson uses Prolog for complex reasoning.  The base notation
for annotating documents in Watson is the XML-based UIMA, which IBM
donated to the Apache project.  For Watson, they implemented a pattern
matching tool that processed UIMA directly.  But they found that Prolog
is far more flexible, general, and *efficient* than the UIMA tools.    (010)

I'm betting on Prolog (and related logic-programming methods) as the
foundation for the next generation of cognitive apps.  If your knowledge
base uses RDF (or JSON), Prolog and related systems are ideally suited
to exploiting anything stored in those representations.  *And* Prolog
is a superset of Datalog, which is better than SQL for RDBMS.    (011)

If you want more evidence, look at AllegroGraph, which is designed to
process RDF-based data, but their native languages are LISP and Prolog.    (012)

RW
> The big need will be in the commercial area where the application will
> be regarded as a competitive advantage not to be shared with anyone.    (013)

It's true that commercial developers won't share anything that they
believe will give them an edge.  But they collaborate on open-source
tools such as Eclipse.  (By the way, Prolog is one of the languages
supported by the Eclipse tools.)    (014)

SS
> AWS is the ecosystem to beat and lord knows, semantic search is
> not Amazon's competitive advantages.    (015)

I agree with both of those points.  Amazon devoted over a dozen years
to designing a robust and efficient cloud that could support a very
large enterprise (namely, their own).    (016)

But the Watson functionality is orthogonal to the cloud services.
I'm sure that IBM's consulting services will (for a price) configure
Watson to be used with any platform anybody is willing to pay for.    (017)

RW
> The decision to open source or go proprietary will depend on whether
> the benefits are viewed as a potential competitive advantage...    (018)

Any large system has a mixture of components at many different levels.
Every business will keep their own data proprietary.  They will also
keep much of their own ontology proprietary.  But any vendor that
does not have a monopoly will collaborate on standard interfaces.    (019)

We must support heterogeneous systems and legacy systems.  Proprietary
ontologies are not a problem, if we have open interfaces.    (020)

John    (021)

_________________________________________________________________
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    (022)

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