AlphaGo was initially trained with data from 100,000 games played by strong amateur players downloaded from the internet. Then, through self-playing and reinforcement learning, it played against different versions of itself millions of times, learning from its errors.
This documentary chronicles DeepMind's AlphaGo project, culminating in its groundbreaking matches against professional Go players. The film explores the complexity of Go, the development of AlphaGo's AI, and the emotional impact of the human-versus-machine competition on both the players and the developers.