AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
The Evolution of AlphaGo: From Expert Knowledge to Self-Play
AlphaGo and its successor, AlphaZero, showcased the power of general methods that can be scaled up using computing infrastructure. AlphaGo relied on prior game knowledge, while AlphaZero learned solely through self play, surpassing human expertise. This development uncoupled the system from human knowledge repositories, allowing it to uncover novel moves that even experts initially considered mistakes. The scalability of these methods, along with advancements in computing power, opens up possibilities for algorithmic innovation and contribution to human knowledge.