Re: Curriculum for AI

From: Michael Roy Ames (
Date: Mon Dec 30 2002 - 13:04:58 MST


You wrote:
> I strongly believe that during AGI proof-of-concept development, it is
> crucial that we not only look at what the system accomplishes, but
> *how* it does it - i.e. the design's construct validity.

I strongly agree.

> [snip] I think it is also important
> for early systems/tests to both deal with noisy environments and
> sense data far in excess of what the system can handle.

For the first four semesters in the curriculum I have specified zero
noise for the teaching interface. This is pragmatism on my part,
because progamming an AGI to deal with noisy data is more work. However
there is no reason the exact same teaching output couldn't have noise
introduced, if this was deemed useful.

In semester 5 (currently very incomplete) the nature of the interface
would inevitably contain noise. The video and audio feeds especially
so. I have plans to create a number of games to learn about dealing
with noise in Semester 5. Throughout the curriculum my purpose is to
present useful ideas, procedures and information to the AGI so that it
can expand its areas of competancy. IMO dealing with noisy environments
is just one of many areas of competancy, all be it a very important one!

> One of the really difficult issues is designing these development
> programs without embodying too much of the designer's specific AGI
> theory/ architecture.

Indeed. This difficulty has caused some revision already when I asking
myself, "How would this game work with a2i2, Novamente, a future Cyc or
FAI seed?" I have described the microdomains digitally for simplicity
of presentation. Many of them could just as easily be presented
visually, or auditorily if that is required.

Michael Roy Ames

This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:41 MDT