Learn more
Join Dexa

AlphaStar Development Journey

The conversation delves into the origins of the AlphaStar project, highlighting the pivotal moment when DeepMind was tasked with tackling StarCraft after the success of AlphaGo. Oriol reflects on the evolution from rule-based strategies to deep reinforcement learning, emphasizing the project's focus on learning behavior rather than pre-defined tactics. As they navigated the complexities of StarCraft, the team faced the challenge of bridging the gap between simpler games and this intricate real-time strategy game.