RE: [WNET] Plan for the next telecon WNET discussion

At 14:58 +0100 13-05-2004, McBride, Brian wrote:
>
>
>>  -----Original Message-----
>>  From: public-swbp-wg-request@w3.org
>>  [mailto:public-swbp-wg-request@w3.org] On Behalf Of Aldo Gangemi
>>  Sent: Thursday, May 13, 2004 2:29 PM
>>  To: public-swbp-wg@w3.org
>>  Subject: [WNET] Plan for the next telecon WNET discussion
>>
>  >
>>  As decided in the last telecon, I propose the following issue to be
>>  discussed with Christiane Fellbaum in the May 27th telecon (30
>>  minutes):
>
>As a matter of form, it is helpful to readers if writers would include
>appropriate links and quotes when reference to previous decisions/discussion
>are made.
>
>Referring to
>
>   http://lists.w3.org/Archives/Public/public-swbp-wg/2004May/0002.html
>
>I see
>
>[[
>ACTION: Aldo to discuss a plan for the May 27th discussion at the next
>Telecon
>]]

OK thanks, doing exact references for everything is always a good practice :)

>  >
>>  (a) description of the actions for short/medium term to be
>>  followed by WNET-TF
>>	(a1) port of Princeton WordNet as an RDF data structure with
>>  an explicit OWL datamodel
>>	(a2) set of recommendations and/or patterns w/alternatives to
>>  reengineer wordnets as OWL ontologies
>
>Hmm, I'd rather separate these out, i.e.
>
>Issue: Encoding WordNet 1.6 in RDF/Owl
>   - Can we have some help from Christiane.  Active involvement would be
>good.  Willingness to review important.

Yes

>   - namespace to use

Yes

>   - Technical issues to discuss e.g. are synsets classes and if so what are
>their instances.

As an anticipation of the report, and to answer 
some of the issue you have raised sometimes ago, 
look at: 
http://www.cogsci.princeton.edu/~wn/index.shtml, 
you'll find in the first paragraph:

WordNet® is an online lexical reference system 
whose design is inspired by current 
psycholinguistic theories of human lexical 
memory. English nouns, verbs, adjectives and 
adverbs are organized into synonym sets, each 
representing one underlying lexical concept. 
Different relations link the synonym sets.

You are served: synsets are *sets* of *synonym 
lexical entries*. Synsets *represent* *one* 
underlying lexical *concept*.

Hence:

1) from the terminological viewpoint, a synset 
can be represented as an owl:Class whose 
instances are lexical entries (words, terms)

2) from the conceptual viewpoint, a synset can be 
represented as an owl:Class whose instances are 
concept's instances (where the concepts are only 
those accepted in the lexicon at hand).

3) A concept instance is *usually* interpreted as 
an entity in the domain of reference of an 
ontology, which is identified by the properties 
of the concept. Of course, there are even 
alternative semantic interpretations for concept 
instances, but the one mostly accepted in 
ontology engineering is the one I have mentioned.

4) Although not considered by WordNet, we can 
also conceive of representing synsets as 
owl:Individual, treating WordNet as a legacy 
database.

Besides my Lab's site, you can look at a lot of 
other work on these issues, for example:

http://www.bultreebank.org/OntoLex02Proceedings.html
http://conferences.atala.org/conferences/fiches/lrec2004.html
http://www.lrec-conf.org/lrec2004/doc/ws/prg_ontolex.pdf

>   - Structuring the data - view files or knowledge server (not sure I
>understand that one - are we thinking they might ship knowledge server code
>as part of the distribution?)

No, Philippe was suggesting that owl-encoded 
wordnets should work on the Semantic Web through 
dedicated knowledge servers rather than as 
scattered views of monolithic files.

>   - Subject to Christiane being satisfied with the results, would he be
>willing ot include in the standard Wnet distribution.

Perfect. BTW ... she be willing ...

>
>Other than mentioning as possible future work, I suggest we don't plan to
>spend any time discussing re-engineering Wordnet or wordnets as ontologies.

Your suggestion clashes against the willing of 
many actors in the scene, besides the intended 
scope of Princeton WordNet. In this committee 
there are obviously persons not interested in 
that kind of work, but I understand they are not 
the totality, for example:

http://lists.w3.org/Archives/Public/public-swbp-wg/2004Apr/0211.html
http://lists.w3.org/Archives/Public/public-swbp-wg/2004May/0023.html.

Moreover, I do not see why a meaningful area of 
investigation, supported by many individuals and 
groups in the Semantic Web and elsewhere, should 
not be addressed by this task force. As a good 
matter of form, look at the discussions in the 
[OEP] task force: none has ever tried to mute or 
to suggest to ignore a position just because she 
is not available to discuss it.

>I think it would be great to have a strawman distributed before to discuss
>at the telecon.  This might be based on Guus' proposal.

It is part of the report due before the telecon.

>I'll be interested to see whether Owl is required for this.  As usual I'm
>inclined to prefer to have a non-Owl version available.

OWL should be used for the datamodel in the 
Stoffel/Guus proposal. Probably the expressivity 
required can be available in RDFS as well.

>  >
>>  (b) availability of Princeton to include an OWL-RDF (datamodel+data)
>>  version (as at (a1)) in the official distribution
>>
>>  (c) availability of Princeton (and/or other initiatives in which
>>  Christiane is involved) to include (possibly by participating in the
>>  development of) our recommendations and patterns at (a2)) in some
>>  official set of guidelines or joint programme
>>
>>  If I receive no further suggestion,
>
>Sorry, but you've got another suggestion.  I feat the task force leader's
>enthusiams and mine lie in different places.

Of course, this is a natural outcome of group 
working ;). I won't force you to discuss *all* 
the issues, you won't force me to shut up. No 
fear, no problem.

>I'll consider these issues as a
>>  reference to produce a report on WNET activity some days before the
>>  telecon, so that we will not need much time to clarify the respective
>>  backgrounds.
>
>Showing willing by having a first cut would seem like a good thing.
>
>Naively
>Brian

OK
Ciao
Aldo
-- 



*;*;*;*;*;*;*;*;*;*;*;*;*;*;*;*;*;*;*;*;*;*
Aldo Gangemi
Research Scientist
Laboratory for Applied Ontology, ISTC-CNR
Institute of Cognitive Sciences and Technologies
(Laboratorio di Ontologia Applicata,
Istituto di Scienze e Tecnologie della Cognizione,
Consiglio Nazionale delle Ricerche)
Viale Marx 15, 00137
Roma Italy
+3906.86090249
+3906.824737 (fax)
mailto://a.gangemi@istc.cnr.it
mailto://gangemi@acm.org
http://www.loa-cnr.it