RE: Curriculum for AI

From: Ben Goertzel (
Date: Mon Dec 30 2002 - 13:15:25 MST

> IRT: "...basal arithmetic..." Yes, that one was already on my list of
> games :) but it would *have* to come *after* the learning of symbols
> (lesson 16). I decided to cut out a *lot* of possible games by
> introducing a new microdomain right after lesson 16. My reason for this
> was to expand the learning space in a new direction, a new dimension.
> Too much learning within a single microdomain would lead to very narrow
> understanding. Beyond a certain point there are diminishing returns
> within a given microdomain.

Hmmm. Well, I think that after getting Novamente past lesson 16, I'd want
to play around with some other games in the purely algorithmic-numerics
domain, before shifting domains. But that's just a matter of taste... I do
understand your point.

> There is a lot more work to do with the doc, and even more fun afterward
> creating an automated teacher and all the game data.

Having automated teachers for these various tests would be great. That
would be a significant contribution to AGI work, I think.

I can already see some controversies arising, however! For instance, in a
genetic programming approach, one can use a loop as a basic given construct.
If one does this, then learning looping will be a lot easier than in a GP
system that is not supplied with such a construct. Not until you get to
really complex tests in richer and more flexible microworlds, will you be at
a stage where simple combinations of hard-wiring and machine-learning start
to fail.

-- Ben G

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