AlphaStar StarCraft II — main technical innovations.
hardAnswer
- Vinyals et al. 2019.
- (1) LSTM policy + transformer over units.
- (2) SL from replays → RL via league play.
- (3) Auto-regressive action distribution over unit + action + coordinates.
- (4) Distributed at scale (thousands of parallel games).
- (5) League: Main / Main Exploiter / League Exploiter roles prevent strategy cycling.
- Grandmaster level; imperfect information + real-time + long horizon.
Check yourself — multiple choice
- Random
- LSTM + transformer over units + SL from replays → RL league play (Main / Main Exploiter / League Exploiter) + auto-regressive action + distributed scale; Grandmaster level
- Same as AlphaGo
- Not real
AlphaStar: LSTM + transformer + SL→league RL; Grandmaster StarCraft II.
#applications#multi-agent
Practise Reinforcement Learning
214 interview questions in this topic.