Removing the target network from deep Q-networks with the mellowmax operator
Publication
, Conference
Kim, S; Asadi, K; Littman, M; Konidaris, G
Published in: Proceedings of the International Joint Conference on Autonomous Agents and Multiagent Systems, AAMAS
January 1, 2019
Deep Q-Network (DQX) is a learning algorithm that achieves humanlevel performance in high-dimensional domains like Atari games. We propose that using an softmax operator, Mellowmax, in DQN reduces its need for a separate target network, which is otherwise necessary to stabilize learning. We empirically show that, in the absence of a target network, the combination of Mellowmax and DQN outperforms DQN alone.
Duke Scholars
Published In
Proceedings of the International Joint Conference on Autonomous Agents and Multiagent Systems, AAMAS
EISSN
1558-2914
ISSN
1548-8403
ISBN
9781510892002
Publication Date
January 1, 2019
Volume
4
Start / End Page
2060 / 2062
Citation
APA
Chicago
ICMJE
MLA
NLM
Kim, S., Asadi, K., Littman, M., & Konidaris, G. (2019). Removing the target network from deep Q-networks with the mellowmax operator. In Proceedings of the International Joint Conference on Autonomous Agents and Multiagent Systems, AAMAS (Vol. 4, pp. 2060–2062).
Kim, S., K. Asadi, M. Littman, and G. Konidaris. “Removing the target network from deep Q-networks with the mellowmax operator.” In Proceedings of the International Joint Conference on Autonomous Agents and Multiagent Systems, AAMAS, 4:2060–62, 2019.
Kim S, Asadi K, Littman M, Konidaris G. Removing the target network from deep Q-networks with the mellowmax operator. In: Proceedings of the International Joint Conference on Autonomous Agents and Multiagent Systems, AAMAS. 2019. p. 2060–2.
Kim, S., et al. “Removing the target network from deep Q-networks with the mellowmax operator.” Proceedings of the International Joint Conference on Autonomous Agents and Multiagent Systems, AAMAS, vol. 4, 2019, pp. 2060–62.
Kim S, Asadi K, Littman M, Konidaris G. Removing the target network from deep Q-networks with the mellowmax operator. Proceedings of the International Joint Conference on Autonomous Agents and Multiagent Systems, AAMAS. 2019. p. 2060–2062.
Published In
Proceedings of the International Joint Conference on Autonomous Agents and Multiagent Systems, AAMAS
EISSN
1558-2914
ISSN
1548-8403
ISBN
9781510892002
Publication Date
January 1, 2019
Volume
4
Start / End Page
2060 / 2062