Q-learning

1 May 1992

Watkins and Dayan prove Q-learning converges. The algorithm that DQN scaled up twenty years later.