Re: [Fwd: Re: Strong players only?]

[johnc@xxxxxxxxx (John Clarke)]

Hi

> Computer go is definitely not like computer chess because we do know very well
> that brute force will *never* be an issue in computer go : a good go programm

But what is the human brain if not the archetypical "brute force" machine?
Neurons aren't very smart - there are just lots of them.

Almost certainly humans use a massively parallel search to choose their moves.

Of course we both agree that humans only evaluate high depth trees to *check*
possible plays, and even those trees are severely pruned.

But the basic algorithm is surely just a very big, very parallel search.  And
the more search the better.

> we have to know how human are so
> good at pattern recognition, image recognition, etc..

Massive, massive parallel processing.  The human visual system is equivalent
to something in the range 1e4 - 1e5 400MHz Pentiums all working together.

Remember our best estimate is that Kasparov embodied between 2 - 10 times
as many chess mips as Deep Blue. ( I estimated that myself, and then was
intrigued --- and very pleased --- to see that Hans Moravec agreed:

 http://frc.ri.cmu.edu/~hpm/hpm.pubs.html 
 
)

Go programming isn't magic, and the solution, when it comes, will involve
*lots* of computation and probably disappointingly little cleverness.
Some cleverness of course, but just enough, and not lots.

ObStrongPlayer:  I guess I expect that a world champion program will be
written by strong programmers, with the added hack of strong go experts
for fine tuning.

Regards
John
johnc@xxxxxxxxxxxxxxxxx

PS.  Sorry for the hectoring tone:  I'm an engineer and consequently a
great fan of brute force since it works so well.