Cecil O. Lynch, MD, MS wrote:
> Hi Kingsley,
>
> I am talking about reasoning. The data sets are HL7 messages passed on
> disease surveillance to the US CDC. The system that we built processes
> an XML message, reasons against the syntax of the message to confirm
> the structure, reasons over the data contained in the message,
> to validate the terminology, then reasons against the content to
> determine the disease status (drug resistance pattern, treatment
> guideline adherence).
Okay, lets take a look at Web Scale Reasoning. (01)
Scenario: Information about Michael Jackson
Data Sources: various data spaces on the Web e.g., Last.FM, Discogs.org,
MusicBrainz, DBpedia, blog posts from Live Journal etc..
Virtuoso instance information: 7.5+ Billion Triples, SPARQL endpoint
live on the Web at: http://lod.openlinksw.com (02)
Actions:
All of this data gets into the Virtuoso Quad store via a variety of
means (remember, Virtuoso is an RDMS and Quad Store Hybrid)
Inference Rules in play: explicit co-reference via "owl:sameAs" and much
fuzzier indirect co-reference using a local axiom that designates
"foaf:name" as being an Inverse Functional Property (IFP). (03)
Links: (04)
1. http://bit.ly/38Jlw4 - Page showing the effect of explicit and
fuzzier reasoning (note the obvious stick out errors re. entity Michael
Jackson); here de-referencing each URI will expose the same expanded
union of data from across all the data spaces (named graphs)
2. http://bit.ly/LKtnt - Page showing same data but without the
inference rules context in play; here each URI will de-reference to its
data space specific data (no union expansion)
> All 50 states send messages to the system. When you start getting
> into the complex reasoning for these types of medical messages for an
> entire country, OWL does not scale.
Really? It doesn't scale if forward-chaining is in play and
unconstrained. I am demonstrating Web Scale reasoning using
backward-chaining. Naturally, since some of the data gets into the Quad
Store via our Sponger Middleware (an RDFizer middleware component),
there is some constrained forward-chaining as part of the sponger
processing pipeline. (05)
Conclusion: (06)
We shouldn't write-off anything, its about using the best combination of
tools for the problem at hand. In this case, the trick is to combine
technology and techniques from a range of realms: raw DBMS and Middleware. (07)
Kingsley
>
> Cecil O. Lynch, MD, MS
>
>
>
> -------- Original Message --------
> Subject: [SPAM] Re: [ontolog-forum] Ontology-based database
> integration
> From: Kingsley Idehen <kidehen@xxxxxxxxxxxxxx>
> Date: Fri, October 09, 2009 10:41 am
> To: "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx>
>
> Cecil Lynch wrote:
> > John,
> >
> > I would say that I completely agree with you that for large volume
> > transaction based systems, OWL or RDF are hopeless from a performance
> > perspective.
> Cecil,
>
> What have you tested? How have you arrived at your conclusions?
>
> You are assuming that RDBMS and RDF Graph hybridization (all the way
> down to the engine core) isn't achievable, right?
>
> What performance are we talking about? Instance data queries,
> reasoning
> over the ABox and TBox etc.?
>
> I really think that when we talk about data integration and the
> prowess
> of RDF, OWL, and what can be achieved re. reasoning etc.. Best we
> point
> to actual data available on the Web.
>
> Links:
>
> 1. http://www4.wiwiss.fu-berlin.de/bizer/BerlinSPARQLBenchmark/ --
> SPARQL Benchmark with RDBMS to RDF component
>
> Kingsley
> > That said, I think that OWL 2 is a great language for basing
> > your model to ensure formalism in the development of the domain
> of interest,
> > but then you need to hand it off to another reasoning approach.
> >
> > There is no perfect ONE language, including Prolog, to approach
> most real
> > world, real time reasoning. In our experience, the best
> performance and
> > logic come from assembling a suite of tools that can work
> together in an
> > orchestrated services platform. There are some problems I want to
> address
> > using DL or regression analysis, others with classification and
> still others
> > with backward chaining. Some tools are better for each of these
> and I think
> > the best systems use them in concert.
> >
> > Cecil
> >
> > -----Original Message-----
> > From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx
> > [mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx
> <http://email.secureserver.net/#Compose>] On Behalf Of John F. Sowa
> > Sent: Friday, October 09, 2009 12:27 AM
> > To: edbark@xxxxxxxx
> > Cc: [ontolog-forum]
> > Subject: Re: [ontolog-forum] Ontology-based database integration
> >
> > Ed,
> >
> > I agree with most of your response to my remark, with the exception
> > of one sentence.
> >
> > JFS>> DL is just one of a large number of logic-based technologies
> > >> that produce useful results for certain kinds of problems.
> > >> Unfortunately, people are being forced to use OWL for tasks
> > >> that it was never designed to do. They go through contortions
> > >> that make Perl look like the epitome of structured elegance.
> >
> >
> > EB> I fully agree. Part of that is the silver bullet mentality:
> > > OWL is the best technology available; so whatever contortion you
> > > have to perform to use it is the best you could have done. And
> > > we are both familiar with the software engineer's pride of
> > > accomplishment in building a Rube Goldberg device to solve a
> > > problem that would be a simple application of a technology he
> > > is unfamiliar with. But we have made progress -- it is not
> > > a primitive AI application coded in Fortran anymore.
> >
> > The point that I very strongly disagree with is that "OWL is
> > the best technology available." At VivoMind, we are delighted
> > when our competitors use OWL because we can translate their
> > sources to Prolog and get orders of magnitude improvement over
> > their "native" implementations.
> >
> > But for heavy-duty lifting (gigabytes and terabytes) we would
> > never dream of using RDF or OWL. Those languages are hopelessly
> > inadequate for truly massive volumes of data. Just note that
> > Google doesn't use those languages. They know better.
> >
> > John
> >
> >
> > _________________________________________________________________
> > 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
> <http://email.secureserver.net/#Compose>
> > 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
> <http://email.secureserver.net/#Compose>
> >
> >
> > ------------------------------------------------------------------------
> >
> >
> > _________________________________________________________________
> > 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
> <http://email.secureserver.net/#Compose>
> > 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
> <http://email.secureserver.net/#Compose>
> >
>
>
> --
>
>
> Regards,
>
> Kingsley Idehen Weblog: http://www.openlinksw.com/blog/~kidehen
> <http://www.openlinksw.com/blog/%7Ekidehen>
> President & CEO
> OpenLink Software Web: http://www.openlinksw.com
> <http://www.openlinksw.com/>
>
>
>
>
>
> _________________________________________________________________
> 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
> <http://email.secureserver.net/#Compose>
> 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
> <http://email.secureserver.net/#Compose>
>
> ------------------------------------------------------------------------
>
>
> _________________________________________________________________
> 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
> (08)
-- (09)
Regards, (010)
Kingsley Idehen Weblog: http://www.openlinksw.com/blog/~kidehen
President & CEO
OpenLink Software Web: http://www.openlinksw.com (011)
_________________________________________________________________
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 (012)
|