I apologize for not being able to participate in today's meeting,
because Thursdays are usually bad for me. (01)
But in reading various notes, I realized that there is one very
strong common thread: Ontologies are ideally suited to hybrid
applications, where there is a multiplicity of different kinds
of algorithms -- logic-based, statistical, and procedural --
that all use the common ontology. (02)
Following are some examples from today's thread: (03)
On 2/24/2011 10:11 AM, Bruce Musicus wrote:
> Although OWL 2.0 has added ranges for scalar variables and
> pattern match to strings, it is difficult to do any computation
> within OWL as part of the definition of the signal class. This
> means that we have to augment OWL with other languages
> and rule systems (e.g. JENA, MATLAB, etc.). (04)
Yes, indeed. We have found that to be very true in our applications
at VivoMind. In fact, that's one reason why I have complained about
OWL services -- we use so many different paradigms for reasoning
with the ontology that the services OWL provides are trivial
compared to what we do with our other tools. (05)
On 2/24/2011 9:32 AM, Eirich, Peter wrote:
> I have found ontology to be viewed as a nice-to-have by most
> project managers but not as an essential building block for a project,
> even though ontology could make practical contributions to --
> at the very least -- the formulation of a project's requirements. (06)
Project requirements are an excellent starting place to begin
using an ontology. But at that level, it's really more of a
terminology than an ontology. As the project grows, more use of
reasoning methods are likely to develop, but different applications
of the same ontology will normally use different methods -- logic-
based, statistical, procedural, etc. (07)
On 2/24/2011 3:12 PM, Ron Wheeler wrote:
> This sounds like exactly kind of problem that the technologies used in
> Watson are designed to solve.
> No clear path to an answer; A number of different ways to evaluate the
> incoming signal; The final answer or answers can be ranked by
> probability of truthfulness; You would like the system to improve itself
> as it experiences different situations. (08)
Yes, that is very much like Watson -- and very much *unlike* OWL. (09)
When we talk about logic-based methods, we have to include
rule-based methods, relational DBs, and learning methods.
Watson uses all of those methods -- but they don't use OWL. (010)
And we should always keep Cyc in mind as the largest formal
ontology on the planet. Everything is built around an ontology,
but they have a very wide range of reasoning methods that they
use with that ontology. (011)
Recommendation: The *first* thing we need to sell is the
use of a standardized *terminology* for the enterprise.
Next, we can show how a little more formalism can make the
terminology easier to use, develop, and maintain. (012)
But at every stage, we should emphasize *hybrids* as the
mainstream use for ontologies. The hierarchy can begin with
a loosely structured terminology at the earliest stages of
requirements and planning. As the project evolves, different
applications are likely to require different ways of using
the core ontology. (013)
OWL can be useful for many projects, but it is only one tool
among many, it is a very hard thing to sell to programmers,
and it doesn't work well with other kinds of tools. (014)
John (015)
_________________________________________________________________
Msg Archives: http://ontolog.cim3.net/forum/ontology-summit/
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontology-summit/
Unsubscribe: mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
Community Files: http://ontolog.cim3.net/file/work/OntologySummit2011/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2011
Community Portal: http://ontolog.cim3.net/wiki/ (016)
|