From: Ben Goertzel (email@example.com)
Date: Sat May 04 2002 - 11:41:39 MDT
A comment on the relation between A2I2 and Novamente.
They are both founded on a "self-organizing network" design. They are both
a bit more high-level and abstract than neural network models. They are
both much simpler than Eliezer thinks an AGI should be, and they both rely
too much on self-organization and emergence to fit neatly and naturally into
Eliezer's "Friendly goal system" framework.
However, Novamente tries to incorporate the best aspects of
semantic-net/logic based AI with the best aspects of neural net based AI,
whereas A2I2 is lower-level, and I view it *very loosely* more as "trying to
get something a little more mind-ish and abstract the neural nettish
approach right". Novamente has many more mind-dynamics going on; it's a lot
more complex than A2I2 (which I view as a bad thing, but I intuitively
believe this level of complexity is going to make the task of getting real
general intelligence to emerge a lot easier).
My worry with A2I2 is that, after they build their digital doggie (or mouse,
or roach, or whatever -- obviously there is no exact animal correlate, it's
not that kind of project), *will their structures and dynamics scale* to the
next level, the level of a more general inteligence with abstract thinking
Peter's intuition is that they will. My intuition is that I'm not so sure,
but I think they may have a fighting chance. They may, in my view, have a
lot of rethinking and new designing to do, to make the leap from a digital
sort-of-animal to something at all capable of abstract thought.
All in all I think it's a really cool project with a lot of promise. A hell
of a lot more exciting, and important, than typical AI research which is
aimed at solving some narrow problem. A2I2 is making a system that will be
capable at least of simple animal-level "experiential interactive learning".
And I hope they can push it further...
-- Ben G
> -----Original Message-----
> From: firstname.lastname@example.org [mailto:email@example.com]On Behalf
> Of Eliezer S. Yudkowsky
> Sent: Saturday, May 04, 2002 11:01 AM
> To: SL4
> Subject: Review of A2I2
> In reviewing Peter Voss's A2I2 I'm working with much less
> information than I
> have for Ben Goertzel's Novamente. My review of Novamente is based on
> reading the Novamente manuscript and having extensive conversations with
> Ben; my review of A2I2 is based on seeing part of a demo at the Foresight
> Gathering before I had to leave for the airport (aaargh!), which hopefully
> was enough to give me a glimpse of the underlying architecture,
> plus various
> bits of surrounding information, a few small exchanges with Voss,
> and so on.
> Unlike Ben with Novamente, Peter Voss has not, as far as I know,
> expressed a
> belief that A2I2's current architecture is adequate to general
> or seed AI; Voss's philosophy of AI is that you start by building
> a mouse.
> Okay. Not my personal approach but not an invalid approach either. After
> seeing the demo my reaction was again that A2I2 was simpler than I'd
> expected, but that could just be because it was a demo.
> What I saw in the demo was an AI with a visual grid operating at multiple
> scales of resolution, and a network of nodes in which each node
> a visual pattern. When A2I2 runs across a visual pattern it hasn't seen
> before - a visual pattern below a defined threshold of similarity to any
> existing nodes - it creates a new node. Parts of A2I2 explicitly
> as being absent from the demo include A2I2's ability to issue
> motor actions
> and a reinforcement system that links perceptions to actions. ("Visual
> pattern" and "motor action" means A2I2 is bouncing around in the visual
> environment of a 2D microworld, not that it's a robot.)
> I guess my overall reaction is that this is fine as far as it
> goes, but I'm
> not sure what Voss plans to do with it when it's finished. Build a better
> -- -- -- -- --
> Eliezer S. Yudkowsky http://intelligence.org/
> Research Fellow, Singularity Institute for Artificial Intelligence
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:38 MDT