math discovery

[Sergio Navega]

>
>Hello all,
>
>Does anyone have any comment or know of resources regarding automated
>math discovery
>involving the principles of compression and/or entropy?
>
>I know about Douglas Lenat's AM program but that started with a
>knowledge of sets
>and used lots of heuristics.
>
>I'd like to use genetic algorithms with a fitness function involving
>compression or
>entropy to evolve structures that would probably start out representing
>simple maths
>like addition and subtraction.
>
>Thanks,
>
>Chuck Carlson
>
>

Hi, Chuck,

You may take a look at Ken Haase's thesis:

Invention and Exploration in Discovery
http://www.media.mit.edu/~haase/thesis/thesis.html

He invented Cyrano, a sequel to Lenat's AM. Unfortunately,
the thesis is very badly formated (HTML) in some regions but 
the important parts are legible.

Regards,
Sergio Navega.

P.S: Great list of references, Gerry!