|
Dear
Pat,
See
below.
Regards
Matthew
Matthew,
If mappings between independently
developed upper ontologies could actually be discovered that gives
the same accuracy as using a single upper ontology, that would be true.
My experience is that mapping between differently developed upper ontologies
is extremely difficult to do with precision, and probably not possible to do
accurately without developing a meta-upper ontology that breaks down the
representations in the different upper ontologies to more fundamental concepts
that can be recombined to give the original UO concepts Even that
process is likely to leave some residue of logical contradiction - not because
logical contradiction is necessary to represent different concepts, but
because it arises when no effort is made to coordinate ontology-building and
therefore there are no constraints on the unprovable assumptions that are made
about abstract things like mental entities. While integration
(barring contradictions) via a meta-upper ontology may be
theoretically possible, I have not seen it done, and extrapolate from my own
efforts that it would be so much more time-consuming than just learning a
common UO that it is effectively a practical economic impossibility.
MW: I agree it is
hard, but it is possible. I agree that only having one ontology
would make it cheaper to achieve interoperation, you might find that
the ontology was not good for a lot of applications though. The only other
problem is that it is not possible to get everyone to agree to one ontology,
and it is also not possible to force
them.
Of course, if one can tolerate a lot of
error, and only wants to do some very shallow reasoning that doesn't even
require a description logic, approximate mappings may function just
fine. If anyone thinks that there is virtue in that, by all means do it
that way. I am concerned to get semantic interoperability that can take
advantage of the much higher logical precision that computers can achieve than
people can with their neural mental processes.
But as I said, if that doesn't convince you, let us
find out how close we can get by trying to build the common UO and seeing
where the residual problems are I doubt that any amount of
discussion will answer the question. I only comment here so that the
assertions about the size, content, or utility of a useful common upper
ontology do not pass as though they were accepted gospel.
MW: Well there are a
number of upper ontologies, so if that is what you want, I challenge you to
pick one and use it. If you are not prepared to do so, why do you think
anyone else would be prepared to use one that you approve
of? Alternaively, you are quite at liberty to produce another UO, but
please do not fool yourself into thinking that it is producing THE
UO.
Pat
Patrick Cassidy
MITRE Corporation
260 Industrial
Way
Eatontown, NJ 07724
Mail Stop: MNJE
Phone: 732-578-6340
Cell:
908-565-4053
Fax: 732-578-6012
Email:
pcassidy@xxxxxxxxx
Dear Pat,
I'm afraid this is not true:
> *** But,
BUT, BUT!!!!! ***
> If any
two ontology-driven systems want to work together, that upper ontology,
WHATEVER IT IS, has to be the same.
If you don't have a
common upper ontology you do need to have a mapping between the upper
ontologies you do have. What is true that having a common upper ontology
give efficiency benefits as the number of ontologies you are trying to
integrate rises. At about 4 ontologies a common ontology (at all levels that
are shared) pays off.
Regards
Matthew
Re: the question - does the upper ontology matter
much?
My recollection from talking to Doug Lenat is that
when he says that the upper ontology is not the most important part
-
he means that several different upper ontologies
might serve equally well, and there is no requirement for one specific
upper ontology.
*** But, BUT, BUT!!!!! ***
If any two ontology-driven systems want to work
together, that upper ontology, WHATEVER IT IS, has to be the
same.
I think Cathy's example of 'person' is a good
illustration of why. If a system decides to change its upper
ontology on the fly, it should be prepared for serious
chaos.
So let us not confuse the issue by failing to
recognize the distinction between one group choosing an upper ontology (or
no upper ontology) for its own isolated purposes, and multiple groups
choosing an upper ontology for communication between semantic
systems. These are very different application scenarios, having
different requirements. The upper ontology has an additional
function in providing the "common sense" implicit knowledge that people
need to fall back on more general principles when the specialized
knowledge and pre-existing experience are insufficient to solve a problem
-- but the ONTACWG deliberations have not reached the implementation level
where that function is of direct concern.
Since the ONTACWG is focused mainly on semantic
interoperability, the questions of whether different upper ontologies can
function together **in an interoperability scenario** should not be
ignored or conflated with a system operating independently to solve some
problem.
The AI community discovered a long time ago that most
of the reasoning people do about practical problems depends heavily on
specialized knowledge of the domain of interest. That has been
confirmed many times, and a recent example would be the Halo
project. But these domain-specific pilots invariably
involve systems operating independently, not systems trying to
**interoperate** semantically. In our discussions I hope we can keep
these very important distinctions right at the forefront, since if they
are ignored it will lead to unnecessary confusion about the
role of the upper ontology.
Pat
Patrick Cassidy
MITRE Corporation
260 Industrial
Way
Eatontown, NJ 07724
Mail Stop: MNJE
Phone:
732-578-6340
Cell: 908-565-4053
Fax: 732-578-6012
Email:
pcassidy@xxxxxxxxx
First of all, apologies to
all for the unfortunate header on my previous message – I’m a bit out of
practice posting to lists like these, and I’m receiving this list in
digest form (which actually I’m finding a bit confusing, I might change
that).
Thanks, Chris, for the
reminder about CL, with which I’m familiar. Following Leo – yes, what
was concerning me was issues of inference engine design qua changeable
arities. It just sounded to me like a recipe for chaos (just imagining
the kind of backchaining that might result….!!!) The suggestion was made
that Bundy was envisaging a system involving a metalevel of reasoning
whereby changes to the formalism itself would be applied at that level
following some independent set of heuristics. I wouldn’t want to rule
that out in principle but would love to hear some actual details of
rules which might be applied at that level.
There was a post by Barry
Smith which I would’ve liked to read but it got scrubbed in my digest
due (I think) to being in HTML – I don’t suppose it could be
re-posted?
John I was interested to
read your description of your evolution away from advocating axiom-rich
ontologising towards recommending something much closer to Wordnet,
indeed a system where (if I have it right) the only assertions in the
ontology are purely definitional, concerning the meanings of terms, and
no empirical claims are made. This sounds prima facie plausible.
However, can a sharp distinction be made between these two categories of
claim? I don’t think so. I think that is one of the lessons of Peirce’s
pragmatism. Consider for instance, “Electrons have a negative charge”.
This was an empirical claim at some early stage of atomic theory yet is
now analytically true. “Cats have whiskers” – empirical or
definitional?
Also, when you say that
assertions in any upper ontology are “much less important” than
assertions in lower-middle and domain ontologies I think that this
statement equivocates dangerously on 2 different meanings of “Much less
important”:
i) Much fewer ontologist
working hours are spent on this part of the ontology
(true)
ii) Removing or
changing these assertions breaks or changes less
(false!!)
I can remember quite a few
times when I was working at Cycorp that relatively minor changes made at
the upper level had (rather fascinatingly) unintended consequences for
folks working on specific projects. (To give just one example, ‘person’
(the legal – social entity) was separated out from ‘human being’ (the
species) and various inferences which depended on those two concepts
being munged together then broke….)
Cheers,
Cathy
|
_________________________________________________________________
Message Archives: http://colab.cim3.net/forum/ontac-forum/
To Post: mailto:ontac-forum@xxxxxxxxxxxxxx
Subscribe/Unsubscribe/Config:
http://colab.cim3.net/mailman/listinfo/ontac-forum/
Shared Files: http://colab.cim3.net/file/work/SICoP/ontac/
Community Wiki:
http://colab.cim3.net/cgi-bin/wiki.pl?SICoP/OntologyTaxonomyCoordinatingWG (01)
|