Yes!
We could actually train neural networks with different characteristics.
Let's say we do one loop through the training dataset, and for each position we add a little notch to the winning probabilities for all positions that have an opponent checker on the bar (and maybe even a bigger notch if there's two or more checkers on the bar). Then we do supervised training with this modified trainset. This will hopefully create a more aggressive player that will be more eager to hit loose on checkers, and hopefully create a player with an attacking style.
Then - Let's say we do one loop through the training dataset, and for each position we subtract a little notch to the winning probabilities for all positions that have a blot that can be hit (and maybe even a bigger notch if there's several of it's blot that can be hit). Then we do supervised training with this modified trainset. This will hopefully create a more careful player that will rather create high stacks than playing flexible. Typically seen by beginner players. 4-1 opening roll is then played 13/8, they seldom split backcheckers etc.
Of course I have no idea if this will work or not. But I think I will be able to do something like this. (But not now as I'm leaving for vacation tomorrow morning)
We probably need some interface that can read custom neural networks. I have lost the touch when it comes to GTK coding, but someone may be able to specify.
-Øystein