Add SmartPredictor about eternabrain HOT 4 CLOSED

Rohan, when doing the play out, you could stochastically choose among moves, weighted by the confidence of the neural net. See, e.g., the DeepMind papers on Atari gameplay learning. On Jul 7, 2017, at 12:26 PM, Rohan Koodli <[email protected]<mailto:[email protected]>> wrote: Checks if base prediction already matches existing base in that location. Example: base_sequence = [1,1,1,1] Prediction = [1630.97,1630.88,1630.56,1630.30] argmax(Prediction) == 0 # base A location = 3 if base_sequence[location] == argmax(Prediction): SmartPredictor() # takes 2nd highest probability and uses that as base — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub<#24>, or mute the thread<https://github.com/notifications/unsubscribe-auth/ABmEixBFFrevKiNeun3r7OuEBnOJY-7-ks5sLoaCgaJpZM4ORV5F>.