bammi with better ai

• original game at www.bammi.com

ai techinques tried

deep reinforcement learning + graph convolutional network

• q-learning
• A3C
• policy gradient
• PGQL

tree search

• minimax search
• monte carlo tree search with original ai as tree policy