Configurable Mirror Descent: Towards a Unification of Decision Making

Pengdeng Li, Shuxin Li, Chang Yang, Xinrun Wang, Shuyue Hu, Xiao Huang, Hau Chan, Bo An
Proceedings of the 41st International Conference on Machine Learning, PMLR 235:28164-28203, 2024.

Abstract

Decision-making problems, categorized as single-agent, e.g., Atari, cooperative multi-agent, e.g., Hanabi, competitive multi-agent, e.g., Hold’em poker, and mixed cooperative and competitive, e.g., football, are ubiquitous in the real world. Although various methods have been proposed to address the specific decision-making categories, these methods typically evolve independently and cannot generalize to other categories. Therefore, a fundamental question for decision-making is: Can we develop a single algorithm to tackle ALL categories of decision-making problems? There are several main challenges to address this question: i) different decision-making categories involve different numbers of agents and different relationships between agents, ii) different categories have different solution concepts and evaluation measures, and iii) there lacks a comprehensive benchmark covering all the categories. This work presents a preliminary attempt to address the question with three main contributions. i) We propose the generalized mirror descent (GMD), a generalization of MD variants, which considers multiple historical policies and works with a broader class of Bregman divergences. ii) We propose the configurable mirror descent (CMD) where a meta-controller is introduced to dynamically adjust the hyper-parameters in GMD conditional on the evaluation measures. iii) We construct the GameBench with 15 academic-friendly games across different decision-making categories. Extensive experiments demonstrate that CMD achieves empirically competitive or better outcomes compared to baselines while providing the capability of exploring diverse dimensions of decision making.

Cite this Paper

BibTeX
@InProceedings{pmlr-v235-li24an, title = {Configurable Mirror Descent: Towards a Unification of Decision Making}, author = {Li, Pengdeng and Li, Shuxin and Yang, Chang and Wang, Xinrun and Hu, Shuyue and Huang, Xiao and Chan, Hau and An, Bo}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, pages = {28164--28203}, 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/li24an/li24an.pdf}, url = {https://proceedings.mlr.press/v235/li24an.html}, abstract = {Decision-making problems, categorized as single-agent, e.g., Atari, cooperative multi-agent, e.g., Hanabi, competitive multi-agent, e.g., Hold’em poker, and mixed cooperative and competitive, e.g., football, are ubiquitous in the real world. Although various methods have been proposed to address the specific decision-making categories, these methods typically evolve independently and cannot generalize to other categories. Therefore, a fundamental question for decision-making is: Can we develop a single algorithm to tackle ALL categories of decision-making problems? There are several main challenges to address this question: i) different decision-making categories involve different numbers of agents and different relationships between agents, ii) different categories have different solution concepts and evaluation measures, and iii) there lacks a comprehensive benchmark covering all the categories. This work presents a preliminary attempt to address the question with three main contributions. i) We propose the generalized mirror descent (GMD), a generalization of MD variants, which considers multiple historical policies and works with a broader class of Bregman divergences. ii) We propose the configurable mirror descent (CMD) where a meta-controller is introduced to dynamically adjust the hyper-parameters in GMD conditional on the evaluation measures. iii) We construct the GameBench with 15 academic-friendly games across different decision-making categories. Extensive experiments demonstrate that CMD achieves empirically competitive or better outcomes compared to baselines while providing the capability of exploring diverse dimensions of decision making.} }
Endnote
%0 Conference Paper %T Configurable Mirror Descent: Towards a Unification of Decision Making %A Pengdeng Li %A Shuxin Li %A Chang Yang %A Xinrun Wang %A Shuyue Hu %A Xiao Huang %A Hau Chan %A Bo An %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-li24an %I PMLR %P 28164--28203 %U https://proceedings.mlr.press/v235/li24an.html %V 235 %X Decision-making problems, categorized as single-agent, e.g., Atari, cooperative multi-agent, e.g., Hanabi, competitive multi-agent, e.g., Hold’em poker, and mixed cooperative and competitive, e.g., football, are ubiquitous in the real world. Although various methods have been proposed to address the specific decision-making categories, these methods typically evolve independently and cannot generalize to other categories. Therefore, a fundamental question for decision-making is: Can we develop a single algorithm to tackle ALL categories of decision-making problems? There are several main challenges to address this question: i) different decision-making categories involve different numbers of agents and different relationships between agents, ii) different categories have different solution concepts and evaluation measures, and iii) there lacks a comprehensive benchmark covering all the categories. This work presents a preliminary attempt to address the question with three main contributions. i) We propose the generalized mirror descent (GMD), a generalization of MD variants, which considers multiple historical policies and works with a broader class of Bregman divergences. ii) We propose the configurable mirror descent (CMD) where a meta-controller is introduced to dynamically adjust the hyper-parameters in GMD conditional on the evaluation measures. iii) We construct the GameBench with 15 academic-friendly games across different decision-making categories. Extensive experiments demonstrate that CMD achieves empirically competitive or better outcomes compared to baselines while providing the capability of exploring diverse dimensions of decision making.
APA
Li, P., Li, S., Yang, C., Wang, X., Hu, S., Huang, X., Chan, H. & An, B.. (2024). Configurable Mirror Descent: Towards a Unification of Decision Making. Proceedings of the 41st International Conference on Machine Learning, in Proceedings of Machine Learning Research 235:28164-28203 Available from https://proceedings.mlr.press/v235/li24an.html.

Related Material