Playing a game

with Reinforcement learning

What we'll gonna do

• play some games with Q-learning, Q-network
• DQN
• With Tensorflow, gym

Machine learning

• Supervised learning
• Unsupervised learning
• Reinforcement learning

Supervised learning

• Need input, target
• Learn from difference between prediction and target
• e.g. mnist, classification

Unsupervised learning

• Need input
• Cluster by distance between inputs
• Can't predict outcome
• e.g. clustering

Reinforcement learning

• input: environment, reward, output: action
• Learn from try
• Model free
• e.g. game play, stock trading

Reinforcement Learning

• Q learning
• Neural Network
• DQN: Deep Q Learning

Basic knowledge

for reinforcement learning

• MDP: Markov Decision Process
• Bellman equation
• Dynamic programming

Basic knowledge

for reinforcement learning

• Value, Polish
• Value function, Polish function
• Value iteration, Polish iteration

Basic knowledge

for reinforcement learning

• Tensorflow 
• or other ML framework

Not today

• Simple game
• Simple code

1. Install gym

pip install gym

https://gym.openai.com

2. Run a game

pip install gym

https://gym.openai.com

2. Run a game

cart pole init

2. Run a game

cart pole

3. Furthermore

Join #machine-learning