yeah, I guess I should read that paper. Not having read it, I'll stick to
my original assertion: link-grammars are categorial grammers are more or
less the same thing, just using a different notation.

There are, however, a few tricks we've learned over the years:

-- proposition 3 deals with cycle-free linkages, and its now quite clear
that cycles are strongly desirable for constraining the number of linkages,
so any grammar that is cycle free is going to loose.  Section 4.1.3 notes
this.

-- non-planar cycles are probably going to be particularly important, but
that research has not yet been done aka "landmark transitivity"

-- assigning liklihoods is my current direction of research. Its where the
rubber hits the road. Its what the word-vector people work on, and I think
that LG/CCG offer superior approaches to  flat vectors.

-- repulsivity or mutual repulsion/exclusion is important for
understanding. Given two similar, nearly identical parses, the correct one
is the one that either aligns the most closely with previous parses, or the
one that is most strongly different from previous parses. The middle ground
must always be ceeded as being wrong.  This is not a philosophical
statement, rather, it is a statement about how human minds actually work,
when working with meaning.

Example: if you are looking at a bird dropping on the sidewalk, and someone
says "look over here", they probably do NOT mean "look at this ladybug
crawling on this leaf". They are directing attention at something very
different, rather than at something else that is similar but narrow.

Linas

On Mon, Jul 31, 2017 at 2:58 PM, Alex <[email protected]> wrote:

> Just for reference, there is down-to-earth, but very interesting and
> valuable article about generating CCG lexicon from Link Grammar lexicon
> https://hal.archives-ouvertes.fr/hal-00487053/document . So, now, as a
> fan of CCG, I will start to value efforts in link grammar area.
>
> --
> You received this message because you are subscribed to the Google Groups
> "opencog" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
> To post to this group, send email to [email protected].
> Visit this group at https://groups.google.com/group/opencog.
> To view this discussion on the web visit https://groups.google.com/d/
> msgid/opencog/a4c083d0-4f31-4205-a95a-611d4b3aa333%40googlegroups.com
> <https://groups.google.com/d/msgid/opencog/a4c083d0-4f31-4205-a95a-611d4b3aa333%40googlegroups.com?utm_medium=email&utm_source=footer>
> .
>
> For more options, visit https://groups.google.com/d/optout.
>

-- 
You received this message because you are subscribed to the Google Groups 
"opencog" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
Visit this group at https://groups.google.com/group/opencog.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/opencog/CAHrUA35OJYyzXnby57BR9E1q_EmKQ5cFO1QiOpCmiGDNOeLS%2BQ%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to