Reinforcement Learning and Artificial Intelligence (RLAI)
	RL-Glue

Edited by Leah Hackman

This is the top-level page for the RL-Glue, a standard software protocol for benchmarking and interconnecting reinforcement learning agents and environments.

The ambition of this web page is to provide a concise and comprehensive guide to setting up and using RL-Glue. Readers are encouraged to start a discussion on general issues regarding the RL-Glue by clicking "Extend this page" at the bottom of this page.  Any questions are encouraged, please do not hesitate to be in contact.

• What is RL-Glue?
• Interfaces and Definitions
• Task specification language

• Writing Agents, Environments and Experiments
• Quick Start Guides to Agents/Environments/Experiments
• Running a Learning Experiment over a Network/the Internet
  

• Frequently asked questions
• Discussions Page
  

• Download RL-Glue

• Changes from Version 1.0
• How to Convert Agents/Environments/Experiments to RL-Glue 2.0
  

• The RL-Library

• Related workshops
  
• RL benchmarks and bake-offs (1st NIPS workshop)
• RL benchmarks and bake-offs II (2nd NIPS workshop)
• First Annual NIPS Reinforcement Learning Competition (NIPS'06)
• Reinforcement Learning Competition and Benchmarking Event (ICML'06)

• Related projects and proposals
  
• RLbench by John Langford and Drew Bagnell
• CLsquare by Martin Reidmiller at al.
• Bandit, by Joannès Vermorel
• RIL Toolbox by Neumann Gerhard and Neumann Stephan

• The Old RL-Glue Website