Tic-Tac-Toe Play

A workflow deploying the network trained in the Tic-Tac-Toe Learn workflow to a KNIME Webportal application to enable interactive play.

This is a companion discussion topic for the original entry at https://kni.me/w/JwmYV-QHc1cWF5xK

Hi Corey,

Thanks nice and interesting workflows.
I tried to train further the NN by increasing the number of iterations up to 100 then the resulting NN was the input to another active learning loop where games were set up to 200 and 25 iterations and then as an input of a third loop where games were set up to 500 and 15 iterations. Even first NN plays better I can not tell a major difference with the last two.
On the other hand my 5 years old daughter wins the NN always.
So it looks like the learning is not very effective. Any ideas why a good learning for the NN is so hard to achieve?