NLP - moving to Word grammar (are there tools?), understanding language, universal reference knowledge base

[Alex]

Hi!

Some questions;

* In http://wiki.opencog.org/w/OpenCogPrime:NLP it is said that there are plans for moving from Link Grammar to Word Grammar. It can be right move (from binary dependencies to networked structures), but is it possible? I can not find any Word Grammar tools (there are lot of Link Grammar tools that OpenCog already employes), are there any Word Grammar tools for integration in OpenCog? What about other grammars - e.g. (combinatory) categorical? The best grammar should be selected for the further efforts. Are there some recent works on Grammars? All the books and articles that wikipedia mentions in its pages about Grammars, are some 10 years old.

* How we can express understanding in OpenCog? Every learning, self-modifying system should be able to estimate the fitness of new pieces of knowledge base (that is generated for the modification of the existing system's knowledge base) with the aim - how well the new pieces of the knowledge base extends the system's understanding about itself and about environment, world. I am aware about the AGI 2016 article about understanding, but maybe there are some other ideas.

* Do AGI systems need universal reference knowledge base (KB)? Humans have such KBs - wikipedia, encyclopedia, canons, universal practices, ideas about what is good and bad. Maybe we can define understanding as a linke to the notions in universal reference KB. Is notion is in KB and if notion is expressed in the terms of other notions in KB, then this notion can be considered as clearly understood and this notion can be used in the definition in new notions? 

[Ben Goertzel]

On Sat, Mar 11, 2017 at 8:54 PM, Alex <alexand...@gmail.com> wrote:
> * In http://wiki.opencog.org/w/OpenCogPrime:NLP it is said that there are
> plans for moving from Link Grammar to Word Grammar. It can be right move
> (from binary dependencies to networked structures), but is it possible? I
> can not find any Word Grammar tools (there are lot of Link Grammar tools
> that OpenCog already employes), are there any Word Grammar tools for
> integration in OpenCog? What about other grammars - e.g. (combinatory)
> categorical? The best grammar should be selected for the further efforts.
> Are there some recent works on Grammars? All the books and articles that
> wikipedia mentions in its pages about Grammars, are some 10 years old.

We are now implementing grammar learning along the lines sketched in

https://arxiv.org/abs/1401.3372

This may lead to something word grammar like in some ways...



--
Ben Goertzel, PhD
http://goertzel.org

“Our first mothers and fathers … were endowed with intelligence; they
saw and instantly they could see far … they succeeded in knowing all
that there is in the world. When they looked, instantly they saw all
around them, and they contemplated in turn the arch of heaven and the
round face of the earth. … Great was their wisdom …. They were able to
know all....

But the Creator and the Maker did not hear this with pleasure. … ‘Are
they not by nature simple creatures of our making? Must they also be
gods? … What if they do not reproduce and multiply?’

Then the Heart of Heaven blew mist into their eyes, which clouded
their sight as when a mirror is breathed upon. Their eyes were covered
and they could see only what was close, only that was clear to them.”

— Popol Vuh (holy book of the ancient Mayas)

[Linas Vepstas]

On Sat, Mar 11, 2017 at 8:54 PM, Alex <alexand...@gmail.com> wrote:

Hi!

Some questions;

* In http://wiki.opencog.org/w/OpenCogPrime:NLP it is said that there are plans for moving from Link Grammar to Word Grammar. It can be right move (from binary dependencies to networked structures), but is it possible? I can not find any Word Grammar tools (there are lot of Link Grammar tools that OpenCog already employes), are there any Word Grammar tools for integration in OpenCog? What about other grammars - e.g. (combinatory) categorical? The best grammar should be selected for the further efforts. Are there some recent works on Grammars? All the books and articles that wikipedia mentions in its pages about Grammars, are some 10 years old.

There's nothing ovewhelmingly new that has happened in the last 10 years; the core theory has been stable, although you can get a better understanding by reading widely.
 

* How we can express understanding in OpenCog? Every learning, self-modifying system should be able to estimate the fitness of new pieces of knowledge base (that is generated for the modification of the existing system's knowledge base) with the aim - how well the new pieces of the knowledge base extends the system's understanding about itself and about environment, world. I am aware about the AGI 2016 article about understanding, but maybe there are some other ideas.

EvaluationLinks.  Ben has multiple books on this topic.
 

* Do AGI systems need universal reference knowledge base (KB)? Humans have such KBs - wikipedia, encyclopedia, canons, universal practices, ideas about what is good and bad. Maybe we can define understanding as a linke to the notions in universal reference KB. Is notion is in KB and if notion is expressed in the terms of other notions in KB, then this notion can be considered as clearly understood and this notion can be used in the definition in new notions? 

the atomspace is designed to hold a KB -- and teh goal is to use automated means to fill it.

--linas

--
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 opencog+unsubscribe@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.
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/0246ab1d-e972-46ff-afaa-fe633ed9d4cc%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Amirouche Boubekki]

There is this pdf that present the field http://www.cs.cmu.edu/~nasmith/slides/sp4nlp.icml09.pdf