there must be a better way

Neural Networks...
Post Reply
ddyer
Posts: 28
Joined: Tue Feb 12, 2008 7:35 pm
Location: Los Angeles

there must be a better way

Post by ddyer » Sun Jan 28, 2018 1:34 am

Regarding using neural nets for games. Doesn't it seem odd that we take a domain
of perfectly discreet objects - black stones, white stones, squares on the board, and
re-represent it as floating point numbers with thousands of parameters?

It seems like there ought to be an approach that uses similar ideas for training and
evaluating neural nets, but uses discreet functions to define activation levels.
my game site is http://Boardspace.net

Rémi Coulom
Posts: 111
Joined: Tue Feb 12, 2008 8:31 pm
Contact:

Re: there must be a better way

Post by Rémi Coulom » Sun Jan 28, 2018 12:02 pm

You can use Binarized Neural Networks:
https://arxiv.org/abs/1602.02830

Post Reply