Multi Agent Reinforcement Learning: AlphaStar League Training
Details
Last meeting, we reviewed the initial announcement of the DeepMind agent AlphaStar and its preliminary competitive results against professional StarCraft 2 players. This time we will focus more on material in section Chapter 9.9.3 of Multi-Agent Reinforcement Learning: Foundations and Modern Approaches which discusses the league training approach to developing robust strategies. We will also look at the associated blog post and Nature paper that explain the training in detail with ablations that show which components of the population training were necessary to reach top performance.
Meetup Links:
Recordings of Previous RL Meetings
Recordings of Previous MARL Meetings
Short RL Tutorials
My exercise solutions and chapter notes for Sutton-Barto
My MARL repository
Kickoff Slides which contain other links
MARL Kickoff Slides
MARL Links:
Multi-Agent Reinforcement Learning: Foundations and Modern Approaches
MARL Summer Course Videos
MARL Slides
Sutton and Barto Links:
Reinforcement Learning: An Introduction by Richard S. Sutton and Andrew G. Barto
Video lectures from a similar course
