We developed an advice model framework to provide theoretical and practical analysis for agents to teach humans and agents in sequential reinforcement learning tasks. The teacher agents assist the students (humans or agents) with action advice when the teachers observe the students reach some critical states. Assuming the teachers are optimal, the students will follow the action advice to achieve better performance.
This section provides the code for running experiments described in (Taylor et al., 2014).
To begin, install Blizzard’s StarCraft game in a Windows machine. Then, download and install the Brood War Application Programming Interface (BWAPI). We used version 3.7.4. Follow the BWAPI install instructions. Once it, and StarCraft injection are both working, install our StarCraft code (C++). Note that there is currently minimal documentation. We hope to have more extensive documentation and installation instructions in the near future.
Our code is based on the Ms. Pac-Man vs. Ghosts contest. You should be able to run this code on any machine that can run Java. Our Pac-Man Java code contains minimial documentation. We hope to have more extensive documentation and installation instructions in the near future.
This software is provided as-is, with no express or implied warranty. If you find it useful, consider dropping us an email, and please cite the above paper.