Leo Obrst asked:
>>>Pat and founding members: care to restate the GOALS we want to
ACHIEVE? (01)
Sure. (02)
The ONTACWG had several specific objectives, all related to the task of
finding relations among the different Knowledge Classification Systems
(KCSs, ontologies, taxonomies, thesauri) that are being built by
different communities. As a working group of the Semantic
Interoperability Community of Practice, our focus is on finding ways to
improve the ability to transfer knowledge between communities that have
their own knowledge classifications suited to their community purposes. (03)
The specific efforts, particularly on the COSMO, slowed down
considerably when preparations were underway for the Upper Ontology
Summit. I hope we can resume soon. I plan to resume discussions on
the COSMO, after first proposing that we consider maintaining a
Controlled Defining Vocabulary of English words to use for creating the
definitions in our KCSs. More on that below. (04)
The list of ONTACWG objectives, from the ONTACWG home page
(http://colab.cim3.net/cgi-bin/wiki.pl?OntologyTaxonomyCoordinatingWG),
are:
(1) To improve the efficiency of creation of knowledge classifications
and semantic knowledge bases by providing a convenient source for
information about ongoing efforts at creating ontologies, taxonomies,
thesauri, graphical knowledge representations, and terminological
resources.
(2) To maintain a list of resources and tools that can be used to
create and effectively use knowledge classifications and knowledge
bases.
(3) To increase the ability to accurately share and reuse knowledge by
building a common upper and mid-level ontology sufficient to provide a
fundamental conceptual vocabulary that can serve to specify the
meanings of the terms and concepts in the community knowledge
classifications, by combination and extension of the basic concepts.
(4) To specify the meanings of the terms in the knowledge
classifications by use of the basic concepts in the higher-level
ontology.
(5) To relate the terms in each community knowledge classification to
the terms in other classifications.
Resource pages are available and pointers to them can be found on the
home page. I would encourage members to use those pages to add in
references to Knowledge Classification projects that they are engaged
in. These pages are Wiki pages, easily modified, and any members who
believe that their work relates to other knowledge classification work
should find a place on the resource pages (or create a new page) to
describe the relations between different KCSs. If you add links or
create new pages, please let us all know by a post to the ONTACWG
group. (05)
In addition to the Wiki pages, we have had some work on two projects:
(1) developing criteria for the types of metadata that would be
useful in a registry containing KCSs - Dagobert Soergel is the lead.
(2) Developing an ontology to serve as the Common Semantic Model
(COSMO). Pat Cassidy is the lead on this. (06)
The work on the COSMO (see pointers at
http://colab.cim3.net/cgi-bin/wiki.pl?CosmoWG) up to this point has
consisted in providing a fragment of an upper ontology that takes
elements from OpenCyc, SUMO, DOLCE, BFO, and ISO15926. We have had
some discussions on this project, but since the run-up to the Upper
Ontology Summit, this project has been quiet. I know that I have been,
and think that others have been, very much distracted by our full-time
jobs recently. I hope we can resume soon. At the UOS we gained some
more perspective on the way the custodians of the existing upper
ontologies view the prospects and methods for finding relations among
their different ontologies, so as to permit re-use of knowledge between
systems using the different upper ontologies. The UOS showed that the
custodians were willing to work toward finding relations among their
ontologies, but developing translation methods will take considerable
effort and will require funding. We hope that one or more of the UO
custodians can find a source of funding for the effort. Until then, I
think that continuing the effort to develop a COSMO based on the
maximum commonality among the existing UOs can provide a useful
experimental resource for testing interoperability via a COSMO in real
applications, (07)
**** New Project: Controlled Defining vocabulary **** (08)
Before resuming work on the COSMO itself, I have made some effort
lately to develop an expanded version of an English "Controlled
Defining Vocabulary", starting from the Longman's controlled
vocabulary. The Longman vocabulary list can be seen in our resource
directory
(http://colab.cim3.net/file/work/SICoP/ontac/reference/LDOCE-definingVo
cabularyList.txt). This original vocabulary has about 2000 words, and
was used by Longman's to defined the 64,000 words in their "Longman's
Dictionary of Contemporary English" (LDOCE). In order to make it a
practical tool for use in defining terms in taxonomies and ontologies,
I have expanded the individual terms (to about 9200 now) to include the
lexical morphological variants (plurals, past tenses, etc.) so it can
be used in the manner of a spell-checker. Some additional basic words
were added to handle needs I perceived, relating to the ontology. I
have also built a very simple Java program that will allow a user to
paste a proposed definition in a window and in two clicks, check to see
if any of the words are not in the controlled vocabulary. (09)
The eventual goal of the controlled vocabulary is to have a basic
controlled English vocabulary that is aligned with the COSMO, which
will have the logical equivalents of all of the terms in the basic
controlled vocabulary. Then, by using the basic controlled vocabulary
to define the terms in one's KCS, it should be possible (in principle)
to convert that definition to a logical specification that will use the
same COSMO, and thereby permit the use of knowledge in that KCS by any
reasoning system using the COSMO. Creating a parser that will convert
the English-language definitions (even though controlled) will of
course be a major project. But since the project will reference
exactly the same COSMO, it can be done collaboratively by dispersed
individuals or groups. (010)
I believe that one of the factors that has hindered use of the
existing upper ontologies is that they are not easy to use. A
controlled defining vocabulary could provide a means to permit easy
entry and retrieval of knowledge without learning the details of the
ontology (or the COSMO). Just how feasible this will be is a question
that this working group can help to explore (more by work than
discussion). (011)
When using the basic controlled vocabulary it will be immediately
evident that users will want to use words not in the basic controlled
vocabulary. In that case, any word in a proposed definition that is
not in the basic vocabulary can itself be defined by use of the basic
vocabulary, and added to a "Supplemented Controlled Vocabulary". The
Supplemented vocabulary can then be used to define additional terms,
and the definitions will still be recursively defined by reference to
the basic defining vocabulary (and correspondingly by reference to the
concepts represented in logical format in the COSMO). I envision users
using three different "controlled vocabularies" - (1) the basic
vocabulary, aligned with the COSMO; (2) a Supplemented Vocabulary, also
shared on a public site, and having additional words defined by use of
the basic defining vocabulary; and (3) a community vocabulary, with
words defined by use of the Supplemented Vocabulary (and recursively,
by the basic controlled vocabulary). Initially, it is likely to be a
nuisance to have to frequently define terms using the basic vocabulary,
so as to be able to use those terms in other definitions. But as the
number of terms already defined (directly or indirectly) via the basic
vocabulary increases, the ability to define specialized terms with
existing vocabularies should become the rule. It will be particularly
interesting to discover which needed words cannot be defined directly
or recursively by the existing basic vocabulary -- thereby signaling a
need for additions to the basic vocabulary (and perhaps to the COSMO as
well). (012)
At present, I think this method may be the most feasible means to
provide resources to help relate Knowledge Classifications to each
other, using only contributed volunteer effort. Funded projects could
make more rapid progress on directly addressing formal relations among
upper ontologies. Even if such a funded effort becomes possible, I
would expect that the definitions created using a common controlled
defining vocabularies will be very helpful in creating formal
definitions for the KCS terms, even if fully automatic conversion to
logical form never becomes possible. (013)
Any members of the ONTACWG who think that this route to practical
interoperability is worth exploring could help in several ways. The
simplest is, by creating definitions of domain terms using the basic
defining vocabulary, or using other terms defined by the basic
vocabulary. Helping to determine what kind of COSMO (upper ontology)
will be adequate for all communities to use as a "conceptual defining
vocabulary" is a second task. Then, some can help to develop a parser
that can take a definition using the basic defining vocabulary terms
and create an accurate translation into a logical form aligned directly
with the COSMO. The last task would complete the basic set of tools
required for this approach to be properly tested. The translation may
not be as difficult as it seems, since we would be concerned only with
translating definitions in a restricted vocabulary, and the pressure
for speedy algorithms that can handle terabytes of data will not be
relevant to this task. (014)
I am in the process of getting formal clearances to deal with
intellectual property issues. After that, I will post the vocabulary
and word-check program to the ONTACWG web site. In the interim,
ONTACWG members may wish to discuss their views on the potential
utility of the controlled-vocabulary approach as outline above to help
ease the use of a COSMO. (015)
Pat (016)
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 (017)
_________________________________________________________________
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 (018)
|