he 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 they harm 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.
- Call for Papers: Recursos educativos digitales
- Curso: El Videojuego es cultura. Taller interactivo de análisis de videojuegos
- Call for papers: Special Issue on Culture Games at ACM Journal on Computing and Cultural Heritage
- “La Noria presenta un proyecto que combina videojuegos e interacción con el entorno para la educación en valores sociales”
- Transmedia Design: Critical and Operational Expertise for Today’s Media Galaxy