From: Ben Goertzel (ben@goertzel.org)
Date: Tue Jan 25 2005 - 18:26:57 MST
Hi Stephen --
That sounds very cool!
While I continue to think that Cyc's basic knowledge representation and
semantics are not workable for general AI, I'm sure you can learn a lot by
experimenting with your Cyc Behavior Engine, and I look forward to hearing
what you learn...
-- Ben
> -----Original Message-----
> From: owner-sl4@sl4.org [mailto:owner-sl4@sl4.org]On Behalf Of Stephen
> Reed
> Sent: Tuesday, January 25, 2005 5:59 PM
> To: sl4@sl4.org
> Cc: Stephen Reed
> Subject: Cyc - Was: When does it pay to play (lottery)?
>
>
>
> On Mon, 24 Jan 2005 21:44:13 -0500, Ben Goertzel <ben@goertzel.org> wrote:
>
> > It is obvious that a GP-based optimizer running on less than a
> million PC's
> > (to be conservative) is not gonna take-off, transcend, become
> self-aware,
> > etc.
> >
> > It's obvious that Cyc, SOAR and EURISKO are not going to do so.
>
> My managers at Cycorp have given me the go-ahead to complete the first
> iteration of the Cyc Behavior Engine and put it into production. Unlike
> our traditional approach, the behavior engine will enable Cyc take the
> initiative, seeking to continuously improve its knowledge and behavior.
> Using an elaboration of Cyc's script and Bayesian vocabulary, behavior
> nodes and methods are compiled out on the fly into java classes. This is
> a facility that has the potential for program synthesis and for
> self-modification. The behavior nodes are structured into an 11-level
> command hierarchy according to the principles of James Albus (a tip from
> Ben). The topmost node has the sole goal of being friendly. The
> structure is vastly scalable if the behaviors are partitioned in a
> fashion similar to large human organizations.
>
> This Friday I will give an internal Cycorp presentation describing the
> behavior engine and the missions of the first 100 or so high
> level nodes.
> As the situation develops and as so far as Cycorp and our sponsors permit
> me to comment, I will provide periodic status reports here for your
> critical review.
>
> As I have posted previously, I will be thinking about what few behaviors,
> when implemented, will most facilitate safe progress on the remainder.
>
> Cheers.
> -Steve
>
> --
> ===========================================================
> Stephen L. Reed phone: 512.342.4036
> Cycorp, Suite 100 fax: 512.342.4040
> 3721 Executive Center Drive email: reed@cyc.com
> Austin, TX 78731 web: http://www.cyc.com
> download OpenCyc at http://www.opencyc.org
> ===========================================================
>
>
This archive was generated by hypermail 2.1.5 : Tue Feb 21 2006 - 04:22:52 MST