On Mon, Sep 24, 2012 at 10:37 AM, Matt Mahoney <[email protected]> wrote:
> As far as I know, OpenCog doesn't have any language capabilities
> (yet),

There is no mature language capability -- i.e. you can't hold a
conversation with an OpenCog system yet.

However, there is code that parses sentences and output semantic
relationships derived therefrom into the Atomspace.   This has been
described in published conference and journal papers.

There is also more experimental and limited code that can transform
some Atoms into English sentences.

This code was hooked together in 2008 to carry out simple
conversations with a virtual dog controlled by OpenCog, though this
functionality was disabled when the Opencog motivational system was
overhauled in 2011, and hasn't been re-enabled yet (it will be).

>and there is currently no development work in this direction.

A description of Ruiting Lian's current NLP development work, in the
OpenCog HK team, is here:

http://wiki.opencog.org/w/Link2Atom

> There are several people working on development, but like any large
> software project, a lot of the work is bogged down on fixing bugs,
> porting issues, and updating the documentation.

That's not really true at all, I'm sorry you have that false
impression ;p ....   OpenCog work is difficult and can be slow, but a
small minority of the developers' time is spent on the things you
mention.  I would prefer if you would stop making statements like
that; you're not part of the OpenCog development team and you really
don't know what's going on in that context !!

>They do have a
> traditional rule based parser (RelEx) and sentence generator which I
> don't believe is integrated into AtomSpace. This is probably not what
> you want if you are interested in a scalable design for language
> learning.

Agreed; RelEx is a "stopgap" system created just to have a simple way
of turning sentences into logical Atoms, for experimental purposes.
The shortcomings of this approach have been discussed in publications
and also (in passing) in the PDF on the above-linked wiki page.

> The current work in OpenCog is toward developing intelligent
> characters for games and virtual worlds. This avoids the
> computationally intensive algorithms required for vision, language,
> and robotics.

We are currently planning to use OpenCog to control a Hanson Robokind
humanoid robot, in a project spanning 2013-2014.  We're seeking research
funding for this from a couple sources, and preliminary indications look good
regarding getting some of it.   David Hanson is an active collaborator in this
effort, and we're also in discussions with Mark Tilden (creator of the
Robo Sapien
robot).  It happens that Tilden is based in Hong Kong, and Hanson spends much
of his time here (where the largest cluster of OpenCog work is currently based).

We have been collaborating with Itamar Arel of the University of Tennessee
in Knoxville, regarding customization of his DeSTIN computer vision architecture
for use as OpenCog's visual and auditory cortex.  This is work in progress.


>There has been some discussion of making a distributed
> version of AtomSpace but IMHO there are going to be severe performance
> problems that make scaling to large numbers of processors impractical.

Can you provide detailed technical reasons for this opinion, based on a specific
critique of the proposed design for a distributed Atomspace, which is described
in the PDF attached to the following wiki page? :

http://wiki.opencog.org/w/DistributedAtomspace

Concrete criticisms on your part might help us improve the design.

General criticisms as you've made, are not very useful.

thx
Ben G


-------------------------------------------
AGI
Archives: https://www.listbox.com/member/archive/303/=now
RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-c97d2393
Modify Your Subscription: 
https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-2484a968
Powered by Listbox: http://www.listbox.com

Reply via email to