Automating Snake Game

Github: https://github.com/JPstCode/SnakeAI

This video explores various techniques for teaching computer how to play the Snake game and ultimately complete it. The featured approaches include:

Hamiltonian cycle
A* pathfinding algorithm
Reinforcement learning (A3C)

Inspiration to the project:
Snake AI (Q-learning):    • AI learns to play SNAKE using Reinforcemen...  
A3C    • Advantage Actor-Critic solves 6x6 Snake (R...   &    • Neural Network Learns to Play Snake using ...  

Code for reinforcement learning is heavily based on these tutorials:
Raymond Yuan, 2018, Deep Reinforcement Learning: Playing CartPole through Asynchronous Advantage Actor Critic (A3C) with tf.keras
and eager execution: https://blog.tensorflow.org/2018/07/d...
Playing CartPole with the Actor-Critic method: https://www.tensorflow.org/tutorials/...
Apoorv Nandan, (2013), Implement Actor Critic Method in CartPole environment: https://keras.io/examples/rl/actor_cr...

A3C Graphs:
R. Sutton & A. Barto, (2014-2015), Reinforcement Learning An Introduction, p. 258: https://web.stanford.edu/class/psych2... (Actor-Critic learning schema)
Sarkar, A. (2018). A Brandom-ian view of Reinforcement Learning towards strong-AI. https://arxiv.org/abs/1803.02912 (Asynchronous training)

Music available on Epidemic Sound:
https://www.epidemicsound.com/track/x...