Near-Optimal Reinforcement Learning with Self-Play under Adaptivity Constraints

Dan Qiao, Yu-Xiang Wang
Proceedings of the 41st International Conference on Machine Learning, PMLR 235:41430-41455, 2024.

Abstract

We study the problem of multi-agent reinforcement learning (MARL) with adaptivity constraints — a new problem motivated by real-world applications where deployments of new policies are costly and the number of policy updates must be minimized. For two-player zero-sum Markov Games, we design a (policy) elimination based algorithm that achieves a regret of ˜O(√H3S2ABK), while the batch complexity is only O(H+loglogK). In the above, S denotes the number of states, A,B are the number of actions for the two players respectively, H is the horizon and K is the number of episodes. Furthermore, we prove a batch complexity lower bound Ω(HlogAK+loglogK) for all algorithms with ˜O(√K) regret bound, which matches our upper bound up to logarithmic factors. As a byproduct, our techniques naturally extend to learning bandit games and reward-free MARL within near optimal batch complexity. To the best of our knowledge, these are the first line of results towards understanding MARL with low adaptivity.

Cite this Paper

BibTeX
@InProceedings{pmlr-v235-qiao24b, title = {Near-Optimal Reinforcement Learning with Self-Play under Adaptivity Constraints}, author = {Qiao, Dan and Wang, Yu-Xiang}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, pages = {41430--41455}, year = {2024}, editor = {Salakhutdinov, Ruslan and Kolter, Zico and Heller, Katherine and Weller, Adrian and Oliver, Nuria and Scarlett, Jonathan and Berkenkamp, Felix}, volume = {235}, series = {Proceedings of Machine Learning Research}, month = {21--27 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v235/main/assets/qiao24b/qiao24b.pdf}, url = {https://proceedings.mlr.press/v235/qiao24b.html}, abstract = {We study the problem of multi-agent reinforcement learning (MARL) with adaptivity constraints — a new problem motivated by real-world applications where deployments of new policies are costly and the number of policy updates must be minimized. For two-player zero-sum Markov Games, we design a (policy) elimination based algorithm that achieves a regret of $\widetilde{O}(\sqrt{H^3 S^2 ABK})$, while the batch complexity is only $O(H+\log\log K)$. In the above, $S$ denotes the number of states, $A,B$ are the number of actions for the two players respectively, $H$ is the horizon and $K$ is the number of episodes. Furthermore, we prove a batch complexity lower bound $\Omega(\frac{H}{\log_{A}K}+\log\log K)$ for all algorithms with $\widetilde{O}(\sqrt{K})$ regret bound, which matches our upper bound up to logarithmic factors. As a byproduct, our techniques naturally extend to learning bandit games and reward-free MARL within near optimal batch complexity. To the best of our knowledge, these are the first line of results towards understanding MARL with low adaptivity.} }
Endnote
%0 Conference Paper %T Near-Optimal Reinforcement Learning with Self-Play under Adaptivity Constraints %A Dan Qiao %A Yu-Xiang Wang %B Proceedings of the 41st International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2024 %E Ruslan Salakhutdinov %E Zico Kolter %E Katherine Heller %E Adrian Weller %E Nuria Oliver %E Jonathan Scarlett %E Felix Berkenkamp %F pmlr-v235-qiao24b %I PMLR %P 41430--41455 %U https://proceedings.mlr.press/v235/qiao24b.html %V 235 %X We study the problem of multi-agent reinforcement learning (MARL) with adaptivity constraints — a new problem motivated by real-world applications where deployments of new policies are costly and the number of policy updates must be minimized. For two-player zero-sum Markov Games, we design a (policy) elimination based algorithm that achieves a regret of $\widetilde{O}(\sqrt{H^3 S^2 ABK})$, while the batch complexity is only $O(H+\log\log K)$. In the above, $S$ denotes the number of states, $A,B$ are the number of actions for the two players respectively, $H$ is the horizon and $K$ is the number of episodes. Furthermore, we prove a batch complexity lower bound $\Omega(\frac{H}{\log_{A}K}+\log\log K)$ for all algorithms with $\widetilde{O}(\sqrt{K})$ regret bound, which matches our upper bound up to logarithmic factors. As a byproduct, our techniques naturally extend to learning bandit games and reward-free MARL within near optimal batch complexity. To the best of our knowledge, these are the first line of results towards understanding MARL with low adaptivity.
APA
Qiao, D. & Wang, Y.. (2024). Near-Optimal Reinforcement Learning with Self-Play under Adaptivity Constraints. Proceedings of the 41st International Conference on Machine Learning, in Proceedings of Machine Learning Research 235:41430-41455 Available from https://proceedings.mlr.press/v235/qiao24b.html.

Related Material