Reinforcement learning competition

By | October 16, 2007

The machine learning community will be holding their second annual reinforcement learning (RL) competition designed to test new RL algorithms in a variety of complex domains including a real-time strategy game, soccer competition, Tetris, helicopter simulation and others. Specifically, the competition will include the following events,

Mountain Car: Perhaps the most well-known reinforcement learning benchmark task, in which an agent must learn how to drive an underpowered car up a steep mountain road.

Tetris: The hugely popular video game, in which four-block shapes must be manipulated to form complete lines when they fall.

Helicopter Hovering: A simulator, based on the work of Andrew Ng and collaborators, which requires an agent to learn to control a hovering helicopter.

Keepaway: A challenging task, based on the RoboCup soccer simulator, that requires a team of three robots to maintain possession the ball while two other robots attempt to steal it.

Real-Time Strategy: A game, based on popular real-time strategy games, which poses exciting new challenges for the reinforcement learning community.

Polyathlon: The agent will face a set of potentially unrelated MDPs with minimal task knowledge and no prior training.

The organizers will release the competition software in 2 weeks time while they will begin accepting team results as early as December of 2007. The final results are expected by July 1st, 2008 just in time for the final event during the International Conference on Machine Learning to be held in Helsinki, Finland, July 6-9, 2008.

More information at the reinforcement learning competition website here.