On The Statistical Complexity of Offline Decision-Making

Thanh Nguyen-Tang, Raman Arora
Proceedings of the 41st International Conference on Machine Learning, PMLR 235:37900-37928, 2024.

Abstract

We study the statistical complexity of offline decision-making with function approximation, establishing (near) minimax-optimal rates for stochastic contextual bandits and Markov decision processes. The performance limits are captured by the pseudo-dimension of the (value) function class and a new characterization of the behavior policy that strictly subsumes all the previous notions of data coverage in the offline decision-making literature. In addition, we seek to understand the benefits of using offline data in online decision-making and show nearly minimax-optimal rates in a wide range of regimes.

Cite this Paper

BibTeX
@InProceedings{pmlr-v235-nguyen-tang24a, title = {On The Statistical Complexity of Offline Decision-Making}, author = {Nguyen-Tang, Thanh and Arora, Raman}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, pages = {37900--37928}, 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/nguyen-tang24a/nguyen-tang24a.pdf}, url = {https://proceedings.mlr.press/v235/nguyen-tang24a.html}, abstract = {We study the statistical complexity of offline decision-making with function approximation, establishing (near) minimax-optimal rates for stochastic contextual bandits and Markov decision processes. The performance limits are captured by the pseudo-dimension of the (value) function class and a new characterization of the behavior policy that strictly subsumes all the previous notions of data coverage in the offline decision-making literature. In addition, we seek to understand the benefits of using offline data in online decision-making and show nearly minimax-optimal rates in a wide range of regimes.} }
Endnote
%0 Conference Paper %T On The Statistical Complexity of Offline Decision-Making %A Thanh Nguyen-Tang %A Raman Arora %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-nguyen-tang24a %I PMLR %P 37900--37928 %U https://proceedings.mlr.press/v235/nguyen-tang24a.html %V 235 %X We study the statistical complexity of offline decision-making with function approximation, establishing (near) minimax-optimal rates for stochastic contextual bandits and Markov decision processes. The performance limits are captured by the pseudo-dimension of the (value) function class and a new characterization of the behavior policy that strictly subsumes all the previous notions of data coverage in the offline decision-making literature. In addition, we seek to understand the benefits of using offline data in online decision-making and show nearly minimax-optimal rates in a wide range of regimes.
APA
Nguyen-Tang, T. & Arora, R.. (2024). On The Statistical Complexity of Offline Decision-Making. Proceedings of the 41st International Conference on Machine Learning, in Proceedings of Machine Learning Research 235:37900-37928 Available from https://proceedings.mlr.press/v235/nguyen-tang24a.html.

Related Material