We recently published a paper on deep reinforcement learning with Double Q-learning, demonstrating that Q-learning learns overoptimistic action values when combined with deep neural networks, even on deterministic environments such as Atari video games, and that this can be remedied by using a variant of Double Q-learning. The resulting Double DQN algorithm greatly improves over the performance of the DQN algorithm.
The popular Q-learning algorithm is known to overestimate action values under certain conditions. It was not previously known whether, in practice, such overestimations are common, whether this harms performance, and whether they can generally be prevented. In this paper, we answer all these questions affirmatively. In particular, we first show that the recent DQN algorithm, which combines Q-learning with a deep neural network, suffers from substantial overestimations in some games in the Atari 2600 domain. We then show that the idea behind the Double Q-learning algorithm, which was introduced in a tabular setting, can be generalized to work with large-scale function approximation. We propose a specific adaptation to the DQN algorithm and show that the resulting algorithm not only reduces the observed overestimations, as hypothesized, but that this also leads to much better performance on several games.