Re: [Bug-gnubg] Training neural nets: How does size matter?

[Joern Thyssen]

On Fri, Aug 30, 2002 at 09:56:28AM +0200, Øystein O Johansen wrote
> Hi,
> 
> gnubg uses a race network, a contact network and a crashed network.
> I think these scheme works OK. The crashed network is a not mature
> yet, but there is work in progress. There has also been discussions
> about splitting into more classes. I guess this will be done
> eventually, but we must take one step at the time. Two years ago
> there was also a network called BPG, (Backgame and Prime). There
> was really some problems with this net, so is was removed.

As far as I remember the problem back then was that a position could be
BPG, then turn into contact, and back into BPG. 

There is discontinuities are the boundaries between two neural nets,
i.e., the evaluate the same position differently. For example, that may
lead to the program refusing to break contact, as the contact neural net
overestimates the contact position. As I remeber it the big problem with
the BPG net was huge discontinuities at the boundaries.

The conclusion was that we should avoid "loops". For example, currently
we have:


Contact -----------  Race   ------ Bearoff
    \               /
     \__ Crashed __/

so no loops. Another way to avoid the discontinuities is the meta-pi
scheme, since this will make the evaluations continuous. The price is
that you, in general, have to evaluate two neural nets.

I've suggested several crashed nets, i.e., one for 0
chequers off, one of 1 chequer off, etc. As Joseph indicated we take
make a series of 15x15=225 of these. Similar for contact and race. So we
could have a sequence of 675 neural nets. Of course, it's a trade-off
between the time it takes to train a neural net and the improvement of
it due to the smaller domain a it covers.

Also, we may add more pseudo-inputs to the neural net. For example,
collect a lot of thumb-rules used by expert or world-class players. One
of the simplest rules is: race, chequers back, and threats. These three
inputs are already in the neural net. 

Jørn

-- 
Joern Thyssen, PhD
Vendsysselgade 3, 3., DK-9000 Aalborg, Denmark
+45 9813 2791 (private) / +45 2077 2689 (mobile) / +45 9633 7036 (work)