Deep Reinforcement Learning with Double Q-learning

@inproceedings{Hasselt2016DeepRL,
  title={Deep Reinforcement Learning with Double Q-learning},
  author={Hado van Hasselt and Arthur Guez and David Silver},
  booktitle={AAAI},
  year={2016}
}
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… CONTINUE READING
Highly Influential
This paper has highly influenced 83 other papers. REVIEW HIGHLY INFLUENTIAL CITATIONS
Highly Cited
This paper has 620 citations. REVIEW CITATIONS
Recent Discussions
This paper has been referenced on Twitter 66 times over the past 90 days. VIEW TWEETS
419 Citations
26 References
Similar Papers

Citations

Publications citing this paper.

621 Citations

01002003002015201620172018
Citations per Year
Semantic Scholar estimates that this publication has 621 citations based on the available data.

See our FAQ for additional information.

References

Publications referenced by this paper.

Similar Papers

Loading similar papers…