I still think either a sigmoid reward curve or a curve which starts as n^2 and ends linear are at least worth to think about ...
You are viewing a single comment's thread from:
I still think either a sigmoid reward curve or a curve which starts as n^2 and ends linear are at least worth to think about ...