[Date Prev] [Date Next] [Thread Prev] [Thread Next] Indexes: Main | Date | Thread | Author

Re: [ba-unrev-talk] Ontologies and volunteers


I did a little ruminating on the problem of auto-generating knowledge structures
from ordinary text a while back. My specific intent was to ascertain whether the
conceptual structures of a work written in say 1650C.E. had lost coherence (or
gained a different coherence perhaps) from updating and translation etc.
My thinking led me to the need to be able to autogenerate reference thesauri
from a corpus of the era concerned in order to statistically detect shifts.
Can computers detect synonymous terms automatically?
My thinking on that was no, mostly because I was unable to turn up anything that
said
I could at the time. Anyone out there ever tried it?
Any information much appreciated.    (01)

I just made another sweep on google though, in which I incidentally pulled up
http://www.ifla.org/IV/ifla62/62-qiyz.htm (1996) which contains this
revealing passage on a different but related topic:
"Automatic language processing, i.e. automatic term extraction is taken as core
of the use of natural language in information retrieval. Unlike the sentences in
English, French, German and Russian, there is no separation marks in Chinese
sentences. A Chinese character can be combined with many other Chinese
characters to form words and phrases which are different in meaning. It is
difficult for computer to recognize which is a Chinese character or which is a
word made up of several characters, thus to separate them automatically, and it
is difficult t o draw a distinction exactly between useful word and useless
word. In the retrieval using Chinese natural language directly, therefore, it is
necessary to solve the technique that the words can be separated automatically
from Chinese sentences by computer. This technique is called Chinese word
separation technique. Researches in this field have been made, and many
proposals on term separation hav e been offered in the recent years. Generally
speaking, some of them can meet actual needs, thus have been used in the system.
One of the practical systems is Word Extraction by Component Dictionary. Most of
them, however, are still in the stage of experiments. It is because automatic
Chinese term extraction is difficult, contrary to Euro-American, there are few
keyword indexes created automatica lly by computers and information retrieval
systems on the basis of technique of automatic term extraction in China. It can
be said, however, that it is not too far to solve the problem of automatic
Chinese term extraction. "    (02)

Makes one realise how powerful the human brain is.    (03)

--
Peter    (04)

----- Original Message -----
From: "Gary Richmond" <garyrichmond@rcn.com>
To: <ba-unrev-talk@bootstrap.org>
Sent: Wednesday, August 07, 2002 8:05 PM
Subject: Re: [ba-unrev-talk] Ontologies and volunteers    (05)


> Thank you for reminding me of The Professor and the Madman, which I read
> very favorable reviews of a few years back.  I agree with you that the
> discussion of ontologies would benefit from reflecting on the evolution
> of dictionaries. You wrote:
>
> > They took the position that in relation to the definition of words,
> > there is value in recording and citing "usage" and that there is no
> > "right" definition, because that would kill any living language.
> >
> In addition to "usage," etymologies (which might be seen as a kind of
> Ur-usage) work to achieve a good balance in relation to definition of
> words IMO (and, of course, the OED includes them, though they are not
> its special thrust).  I've been fascinated with etymologies since as I
> boy I discovered such interesting facts as that the Latin root altus
> figures in words meaning both high (e.g., altitude) and low (e.g. alto,
> the lower female voice), and that indeed any number of ontologies point
> to a broader spectrum of original meaning for particular words than much
> of the later usage might suggest.
>
> Indeed the usage of some words have taken us far from their "roots." For
> example, education (from Latin, ducere, to draw, and e- out) which seems
> first to have meant something like the drawing out of the native
> potential of a person,  has since come to mean something closer to
> "putting in." (There is a variant of this etymology that suggests the
> "drawing out" was of a child at its birth by a midwife--but that
> birthing idea seems rather apt for education as well.)
>
> Again, thanks for getting me to think about these matters again,
> something I haven't done for quite some time.
>
> Gary Richmond
>
> Mei Lin Fung wrote:
>
> > Has anyone read the Professor and the Madman? About the making of the
> > Oxford English Dictionary? I found it very inspiring.  Perhaps our
> > discussion of ontologies might reflect on the history of dictionaries?
> > Taking lessons from the evolution of dictionaries as an early
> > knowledge artifact?
> >
> >
> >
> > They took the position that in relation to the definition of words,
> > there is value in recording and citing "usage" and that there is no
> > "right" definition, because that would kill any living language.
> >
> >
> >
> > About the recruitment of many many volunteers to contribute to this
> > massive 44+ year project. That it took 27 years from the first
> > suggestion of a dictionary to record all English words, to the actual
> > beginning of the project,.... Yes, to get to just the beginning! Our
> > human systems do not evolve easily, but inexorably, they work out what
> > needs to work and mysteriously, find ways to do it.
> >
> >
> >
> >
> >
> > Mei Lin Fung
> >
> >
> >
>
>    (06)