AI agent using a modified Q-learning and neural nets for reinforcement learning on grid based games like Candy Crush and Bejeweled. See our approach on our poster.
siddie / candy-crush-rl Goto Github PK
View Code? Open in Web Editor NEWThis project forked from karan1149/candy-crush-rl
Candy Crush agent using a modified Q-learning and neural nets for reinforcement learning.