ontolog-forum
[Top] [All Lists]

Re: [ontolog-forum] semantic distance (was: the god of boundary)

To: "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: Duane Nickull <dnickull@xxxxxxxxx>
Date: Fri, 09 Feb 2007 13:55:10 -0800
Message-id: <C1F22DBE.812C%dnickull@xxxxxxxxx>
In the early days of the UN/CEFACT CCTS work, we realized that this concept
was an interesting yet possibly useless concept, especially in the context
of building a graphical representation of the taxonomies, ontologies into a
metadata registry facility.  The 2 dimensional model was not sufficient
given the CCTS uses an 8 dimensional matrix to overlay context modifiers on
taxonomic nodes as well as a grounding of each node as a specialized child
of the parent ontological topic.  The system has many flaws so I wouldn't
recommend reading up on it but it did raise some very interesting questions.
We had a lot of trouble explaining it given most people have a great deal of
trouble visualizing an 8 dimensional matrix.    (01)

First - "distance" is not a symmetrical relationship nor is it always
transitive.  In fact, the "distance" may be different based on the view.  I
use distance here to denote a close approximation of what I think you are
all talking about.  YAMMV.    (02)

What is the real driver for defining it?  I cannot really see any use cases
that make a great deal of sense although I am more than willing to concede
ignorance.    (03)

Interesting discussion though.    (04)

Duane    (05)




On 2/9/07 1:46 PM, "Steve Newcomb" <srn@xxxxxxxxxxxxx> wrote:    (06)

> FYI.
> 
> For a *completely* different take on "measuring semantic distance",
> see ieml.org , and click on "Dictionnaire".  Although it may not be
> obvious at that website, what Pierre Lévy (Canada Chair for Collective
> Intelligence, U. of Ottawa) has done there is to propose an address
> space within which every kind of human transaction (political,
> spiritual, literary, etc ...) has a unique and well-grounded address.
> The fact that all addresses are expressed in terms of two dialectics,
> one with three poles and one with two poles, is what makes the
> distances in the space "measurable".  (Although the multiple possible
> approaches to the making of actual measurements based on these poles
> is an area of ongoing research.)  It's not yet completely clear to me
> how the grounding actually happens, but I'm convinced that it's
> happening in at least *some* minds.
> 
> Regardless of whether it actually "works" in some particular sense or
> another, it's one of the most intriguing things I've seen.  For
> example, it has forced me to question my long-held conviction that
> there's no comprehensive way to categorize relationships, hyperlinks,
> etc.  Lévy's approach implicitly discards everything that's outside
> the human frame of reference (as if he were acting on the converse of
> Montaigne's observation that "nothing human was alien to him"), and
> explicitly says that, at least *within* the human frame of reference,
> everything is touched, in a knowable, groundable fashion, by his two
> dialectics.
> 
> 
> "Adrian Walker" <adriandwalker@xxxxxxxxx> writes:
> 
>> You wrote...
>> 
>> The lack of capability of computing the distance between two
>> words/phrases/ideas still cripples all current semantic tools.
>> 
>> The replacement for ontology will possess, as its major
>> term/characteristic, the ability to compute semantic distances.
>> 
>> I'd be interested in your thoughts please about the semantic distance
>> proposal in [1].
>> 
>> The basic idea is that, in walking an abstraction hierarchy up from A and
>> back down to B, the semantic distance is measured by ambiguity, that is, the
>> number of wrong turns you could make along the way.
>> 
>> Thanks!                                -- Adrian
>> 
>> [1]
>> www.reengineeringllc.com/Internet_Business_Logic_and_Semantic_Web_Presentatio
>> n.pdf,
>> slides 40-50
>> 
>> Internet Business Logic (R)
>> A Wiki for Executable Open Vocabulary English
>> Online at www.reengineeringllc.com
>>                                 Shared use is free
>> Adrian Walker
>> Reengineering
>> Phone: USA 860 830 2085
>> 
>> 
>> 
>> On 2/8/07, Mike Brenner <mikeb@xxxxxxxxx> wrote:
>>> 
>>> Bob Smith wrote:
>>>> Is this all about the old topic called "Semantic Distance" we played
>>>> with in the late 1960's and early 1970's, but now with a better GIS-OGC
>>>> sense of "Distance"?
>>> 
>>> 
>>> The lack of capability of computing the distance between two
>>> words/phrases/ideas still cripples all current semantic tools.
>>> 
>>> The replacement for ontology will possess, as its major
>>> term/characteristic, the ability to compute semantic distances.
>>> 
>>> Mike
>>> 
>>> 
>>> _________________________________________________________________
>>> 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
>>> 
>>> 
>>  
>> _________________________________________________________________
>> 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
>>      (07)

-- 
**********************************************************
Sr. Technical Evangelist - Adobe Systems, Inc.           *
Chair - OASIS SOA Reference Model Technical Committee    *
Blog: http://technoracle.blogspot.com                    *
Music: http://www.mix2r.com/audio/by/artist/duane_nickull*
**********************************************************    (08)


_________________________________________________________________
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    (09)

<Prev in Thread] Current Thread [Next in Thread>