[Corpora-List] WordNet vs Ontology

[shams]

Dear Prof. Sowa,

let me put two of your points together:
>
> But I also believe that the key to having an open-ended variety of
> specialized ontologies is to make the computers do what people do:
> extend their ontologies automatically by reading books.

> "I don't believe in word senses."


I do agree with the first point. This is what we tried to do in Hasti 
project at 2004.
This point confirms that what machine should extract ontology from is 
text or any thing else people
learn from, such as image, voice, etc. (and ofcourse reasoning over the 
extracted knowledge)
if we consider texts as the source of knowledge extraction for an 
ontology learning (OL)system
then we cannot avoid word senses.
children may learn about a concept even before knowing the word 
referring to it (its name).
because they see it, hear it, smell it or touch it and feel it. but 
machines which learn from text
can do none of these. so what they get from the real world are the 
words.

I also do agree that syntax is not the appropriate way (or the only 
appropriate way) to extract concepts and
conceptual relations from text. But if human is our point of reference 
then ontologies should be extracted or
discovered (almost) from scratch without an initial ontology at first 
and then be extended according to the previous discovered
knowledge.

I believe that at the first stage (learning from scratch) what we need 
is just a small kernel of essential meta knowledge to let the system
discover the knowledge from whatever resource it has.
the essential meta knowledge for a OL system which learns from scratch 
from texts should contains some linguistic knowledge
and this is unavoidable.
in the next stages the system has acquired a set of ontology 
(knowledge) elements (I call them ontels) and now can continue
to extend its ontology automatically. this is the point from which some 
OL systems which use an apriori or initial ontology start from.

Actually cyc assumes that its OL system is like a matured person with 
an ontology extracted some way while Hasti assumes
that its OL system is like a newborn infant whose only interface to the 
world is reading texts.
both of them are somehow similar to what human do. the main question 
here is that in a second type OL system like what Cyc uses or
many others proposed in the literature (starting from an existing 
ontology) is the initial ontology the right point to start from?
I have some points on this topic but it may be irrelevant to this email 
subject.

I would like to have your opinion about the above facts. are our 
believes the same? if so the word senses are important in
understanding the text or discover knowledge from it. isn't it?

Thanks
Merhnoush Shamsfard

On 2014-08-08 08:41, John F Sowa wrote:
> On 8/7/2014 10:57 PM, Ken Litkowski wrote:
>> It would seem to me that our goal should be a classification
>> of all existing things (not to exclude the narrower types).
>
> Yes, but note the slides I suggested in my first note:
>
>    http://www.jfsowa.com/talks/kdptut.pdf
>
> Slides 7 to 9:  Cyc project.  30 years of work (since 1984).
> After the first 25 years, 100 million dollars and 1000 person-years
> of work (one person-millennium!), 600,000 concepts, defined by
> 5,000,000 axioms, organized in 6,000 microtheories -- and counting.
>
> Slide 10:  2300 years of universal ontology schemes -- and counting.
>
>> The Brandeis Shallow Ontology attempts to do this, and incidentally
>> is being used to characterize arguments of verbs in Patrick Hanks
>> corpus pattern analysis, i.e., in the imperfect world of language.
>
> I strongly believe in shallow, underspecified ontologies -- 
> especially
> when they're supplemented with lots of lexical information about 
> verbs
> and their characteristic patterns.
>
> But I also believe that the key to having an open-ended variety of
> specialized ontologies is to make the computers do what people do:
> extend their ontologies automatically by reading books.
>
> Lenat made the mistake of assuming that you need to hand-code
> a huge amount of knowledge before a system can start to read
> by itself.  But that's wrong.  You need to design a system that
> can automatically augment its ontology every step of the way.
>
> John
>
> _______________________________________________
> UNSUBSCRIBE from this page: http://mailman.uib.no/options/corpora
> Corpora mailing list
> Corpora at uib.no
> http://mailman.uib.no/listinfo/corpora


_______________________________________________
UNSUBSCRIBE from this page: http://mailman.uib.no/options/corpora
Corpora mailing list
Corpora at uib.no
http://mailman.uib.no/listinfo/corpora