Reinforcement learning introduction
States, actions, policies, rewards and value.
Topic hub
Machine learning and reinforcement learning
A path connecting data science, reinforcement learning, simulation environments and analysis practices with reproducible examples.
Start with the reinforcement learning introduction to understand states, actions and rewards.
Then move into tabular methods, approximate solutions and environments such as Google Research Football or MuZero.
Tabular solutions
Windy Gridworld and tabular methods.
Deep reinforcement learning
Lunar Lander and neural approximation.
Google Research Football
Football environment for reinforcement agents.
MuZero
Reinforcement learning and planning.