High-Level Strategy Selection under Partial Observability in StarCraft: Brood War
This addresses strategy adaptation for AI bots in adversarial games, but it is incremental as it builds on existing reinforcement learning methods with a specific auxiliary task.
The paper tackles the problem of high-level strategy selection in real-time strategy games under partial observability, using reinforcement learning with an auxiliary prediction task during training, and reports substantial win rate improvements over a fixed-strategy baseline in experiments against strong StarCraft: Brood War bots.
We consider the problem of high-level strategy selection in the adversarial setting of real-time strategy games from a reinforcement learning perspective, where taking an action corresponds to switching to the respective strategy. Here, a good strategy successfully counters the opponent's current and possible future strategies which can only be estimated using partial observations. We investigate whether we can utilize the full game state information during training time (in the form of an auxiliary prediction task) to increase performance. Experiments carried out within a StarCraft: Brood War bot against strong community bots show substantial win rate improvements over a fixed-strategy baseline and encouraging results when learning with the auxiliary task.