Classic control problems from the RL literature.
Learn to imitate computations.
Reach high scores in Atari 2600 games.
Play classic board games against strong opponents.
Continuous control tasks in the Box2D simulator.
Continuous control tasks, running in a fast physics simulator.
Tune parameters of costly experiments to obtain better outcomes.
Simple text environments to get you started.
NChain-v0 (experimental) (by @machinaut)
n-Chain environment This game presents moves along a linear chain of states, with two actions: 0) forward, which moves along the chain but returns no reward 1) backward, which returns to the beginning and has a small reward The end of the chain, however, presents a large reward, and by moving 'forward' at the end of the chain this large reward can be repeated. At each action, there is a small probability that the agent 'slips' and the opposite transition is instead taken. The observed state is the current state in the chain (0 to n-1).
Environments to test various AI safety properties.
Minecraft environments based on Malmo.
PyGame Learning Environment
Reach high scores in games from the PyGame Learning Environment.
Half-field offense Soccer environments
Doom environments based on VizDoom.