Code, Data & Simulators

Code

• https://github.com/keras-rl/keras-rl
• https://github.com/openai/baselines

Demonstration Datasets

• Basketball (sign via Stats)
  • http://www.basketball-ai.com
  • Generative Multi-Agent Behavioral Cloning, by Eric Zhan, Stephan Zheng, Yisong Yue, Long Sha, Patrick Lucey. [gitbub]
  • Generating Long-term Trajectories using Deep Hierarchical Networks, by Stephan Zheng, Yisong Yue, Patrick Lucey. NIPS 2016.
• Soccer (sign via Stats)
  • Coordinated Multi-Agent Imitation Learning, by Hoang Le, Yisong Yue, Peter Carr, Patrick Lucey. ICML 2017. [github]
• Broadcast Camera
  • Smooth Imitation for Online Sequence Prediction, by Hoang Le, Andrew Kang, Yisong Yue, Peter Carr. ICML 2016. [github]
• Animation (ask us)

Simulators

• Contextual Bandit Simulator
• OpenAI Gym:
  • https://gym.openai.com/
  • A wrapper for Atari, Mujoco, and other environments
• Deepmind Labs
  • 3D maze environments
  • https://github.com/deepmind/lab
• VizDoom
  • Based on the classic 3D-shooter game Doom
  • https://github.com/mwydmuch/ViZDoom
• Mujoco:
  • (imperfect) physics simulator
  • http://www.mujoco.org/
• Arcade Learning Environment (for Atari games)
  • https://github.com/mgbellemare/Arcade-Learning-Environment/releases
  • ask Hoang for tips on Atari games, especially Montezuma's Revenge (hmle@caltech.edu)
    • Hoang's setup: https://github.com/hoangminhle/hierarchical_IL_RL
• pyBullet
  • Experimental open-source physics environments, mimics Mujoco
  • https://github.com/benelot/bullet-gym
• Minecraft
  • Project Malmo
  • https://github.com/Microsoft/malmo
• ELF (Extensible Learning Framework)
  • platform for toy real-time strategy games
  • https://github.com/facebookresearch/ELF
  • https://code.facebook.com/posts/132985767285406/introducing-elf-an-extensive-lightweight-and-flexible-platform-for-game-research/
• Pommerman
  • Based on Bomberman
  • Competition ending at 5/25 and NIPS 2018!
  • https://www.pommerman.com/
• Bridge
  • ask Stephan (stephan@caltech.edu) for code
• Poker
  • ask Stephan (stephan@caltech.edu) for code that could be used as a framework
• Learning for A* Search
  • ask Jialin (jssong@caltech.edu)
• Gridworld - multi-room environment
  • https://github.com/hoangminhle/hierarchical_IL_RL
  • ask Hoang (hmle@caltech.edu) for code that could be useful for gridworld