Safe and Robust Subgame Exploitation in Imperfect Information Games

Zhenxing Ge, Zheng Xu, Tianyu Ding, Linjian Meng, Bo An, Wenbin Li, Yang Gao
Proceedings of the 41st International Conference on Machine Learning, PMLR 235:15255-15270, 2024.

Abstract

Opponent exploitation is an important task for players to exploit the weaknesses of others in games. Existing approaches mainly focus on balancing between exploitation and exploitability but are often vulnerable to modeling errors and deceptive adversaries. To address this problem, our paper offers a novel perspective on the safety of opponent exploitation, named Adaptation Safety. This concept leverages the insight that strategies, even those not explicitly aimed at opponent exploitation, may inherently be exploitable due to computational complexities, rendering traditional safety overly rigorous. In contrast, adaptation safety requires that the strategy should not be more exploitable than it would be in scenarios where opponent exploitation is not considered. Building on such adaptation safety, we further propose an Opponent eXploitation Search (OX-Search) framework by incorporating real-time search techniques for efficient online opponent exploitation. Moreover, we provide theoretical analyses to show the adaptation safety and robust exploitation of OX-Search, even with inaccurate opponent models. Empirical evaluations in popular poker games demonstrate OX-Search’s superiority in both exploitability and exploitation compared to previous methods.

Cite this Paper

BibTeX
@InProceedings{pmlr-v235-ge24b, title = {Safe and Robust Subgame Exploitation in Imperfect Information Games}, author = {Ge, Zhenxing and Xu, Zheng and Ding, Tianyu and Meng, Linjian and An, Bo and Li, Wenbin and Gao, Yang}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, pages = {15255--15270}, 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/ge24b/ge24b.pdf}, url = {https://proceedings.mlr.press/v235/ge24b.html}, abstract = {Opponent exploitation is an important task for players to exploit the weaknesses of others in games. Existing approaches mainly focus on balancing between exploitation and exploitability but are often vulnerable to modeling errors and deceptive adversaries. To address this problem, our paper offers a novel perspective on the safety of opponent exploitation, named Adaptation Safety. This concept leverages the insight that strategies, even those not explicitly aimed at opponent exploitation, may inherently be exploitable due to computational complexities, rendering traditional safety overly rigorous. In contrast, adaptation safety requires that the strategy should not be more exploitable than it would be in scenarios where opponent exploitation is not considered. Building on such adaptation safety, we further propose an Opponent eXploitation Search (OX-Search) framework by incorporating real-time search techniques for efficient online opponent exploitation. Moreover, we provide theoretical analyses to show the adaptation safety and robust exploitation of OX-Search, even with inaccurate opponent models. Empirical evaluations in popular poker games demonstrate OX-Search’s superiority in both exploitability and exploitation compared to previous methods.} }
Endnote
%0 Conference Paper %T Safe and Robust Subgame Exploitation in Imperfect Information Games %A Zhenxing Ge %A Zheng Xu %A Tianyu Ding %A Linjian Meng %A Bo An %A Wenbin Li %A Yang Gao %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-ge24b %I PMLR %P 15255--15270 %U https://proceedings.mlr.press/v235/ge24b.html %V 235 %X Opponent exploitation is an important task for players to exploit the weaknesses of others in games. Existing approaches mainly focus on balancing between exploitation and exploitability but are often vulnerable to modeling errors and deceptive adversaries. To address this problem, our paper offers a novel perspective on the safety of opponent exploitation, named Adaptation Safety. This concept leverages the insight that strategies, even those not explicitly aimed at opponent exploitation, may inherently be exploitable due to computational complexities, rendering traditional safety overly rigorous. In contrast, adaptation safety requires that the strategy should not be more exploitable than it would be in scenarios where opponent exploitation is not considered. Building on such adaptation safety, we further propose an Opponent eXploitation Search (OX-Search) framework by incorporating real-time search techniques for efficient online opponent exploitation. Moreover, we provide theoretical analyses to show the adaptation safety and robust exploitation of OX-Search, even with inaccurate opponent models. Empirical evaluations in popular poker games demonstrate OX-Search’s superiority in both exploitability and exploitation compared to previous methods.
APA
Ge, Z., Xu, Z., Ding, T., Meng, L., An, B., Li, W. & Gao, Y.. (2024). Safe and Robust Subgame Exploitation in Imperfect Information Games. Proceedings of the 41st International Conference on Machine Learning, in Proceedings of Machine Learning Research 235:15255-15270 Available from https://proceedings.mlr.press/v235/ge24b.html.

Related Material