Game and AI player training

This release contains a working tictactoe game, a reinforcement training for a player and some player logic implementations.

Main functionality

• two tictactoe games 3x3 and 4x4
• minimax logic player
• simple logic player
• trainable neural network player
• q-learning implementation for reinforcement learning
• persist trained networks and play against them
• UI to play against different players
• write down CSV files from test epoch results to visualize learning charts
• run training with different easy configurable parameters