On Feb 2, 2008 4:45 AM, John F. Sowa <sowa@xxxxxxxxxxx> wrote:
> ...the knowledge needed
> for language understanding is encoded in the surface patterns, and
> much can be done without an intermediate form that resembles logic. (01)
Glad to find you writing this, John. (02)
> > Or talk to a linguist about the many senses of "in" used in
> > English (approximately 30, though it is hard to be exact),
> > which require an ontology to be used in order to disambiguate them.
>
> What I am questioning is the need for an a priori list of all
> possible interpretations. That is not the assumption that
> underlies GAT or the statistical methods of MT. (03)
I agree with this too. It is only an assumption that we need an
ontology to "disambiguate" natural language. (04)
The opposite is just as likely: that we need natural language to
encode the complexity of ontology. (05)
As a quick list of resources for anyone interested in playing with
these ideas you could start with Dekang Lin's demos. Here's one which
displays classes found using raw word proximity: (06)
http://www.cs.ualberta.ca/~lindek/demos/proxysim.htm (07)
There is also a SourceForge project, SenseClusters, for an open source
implementation of some of the algorithms: (08)
http://www.d.umn.edu/~tpederse/senseclusters.html (09)
Ted Pedersen has a video tutorial for SenseClusters available off this page: (010)
http://www.fask.uni-mainz.de/lk/videoarchive/ (011)
For all that, I agree there is a problem deriving ontologies using
these algorithms. The difference is that I don't think the problem is
that we don't find meaningful classes. (012)
Rather I think what has stopped these techniques being used more is
that we find _too many_ meaningful classes, and there is no objective
way to select between them. (013)
And the answer to that I think is in this "surface complexity" of
natural language. (014)
-Rob (015)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Subscribe/Config: 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 Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx (016)
|