Who Needs to Know? Minimal Knowledge for Optimal Coordination

Niklas Lauffer, Ameesh Shah, Micah Carroll, Michael D Dennis, Stuart Russell
Proceedings of the 40th International Conference on Machine Learning, PMLR 202:18599-18613, 2023.

Abstract

To optimally coordinate with others in cooperative games, it is often crucial to have information about one’s collaborators: successful driving requires understanding which side of the road to drive on. However, not every feature of collaborators is strategically relevant: the fine-grained acceleration of drivers may be ignored while maintaining optimal coordination. We show that there is a well-defined dichotomy between strategically relevant and irrelevant information. Moreover, we show that, in dynamic games, this dichotomy has a compact representation that can be efficiently computed via a Bellman backup operator. We apply this algorithm to analyze the strategically relevant information for tasks in both a standard and a partially observable version of the Overcooked environment. Theoretical and empirical results show that our algorithms are significantly more efficient than baselines. Videos are available at https://minknowledge.github.io.

Cite this Paper

BibTeX
@InProceedings{pmlr-v202-lauffer23a, title = {Who Needs to Know? {M}inimal Knowledge for Optimal Coordination}, author = {Lauffer, Niklas and Shah, Ameesh and Carroll, Micah and Dennis, Michael D and Russell, Stuart}, booktitle = {Proceedings of the 40th International Conference on Machine Learning}, pages = {18599--18613}, year = {2023}, editor = {Krause, Andreas and Brunskill, Emma and Cho, Kyunghyun and Engelhardt, Barbara and Sabato, Sivan and Scarlett, Jonathan}, volume = {202}, series = {Proceedings of Machine Learning Research}, month = {23--29 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v202/lauffer23a/lauffer23a.pdf}, url = {https://proceedings.mlr.press/v202/lauffer23a.html}, abstract = {To optimally coordinate with others in cooperative games, it is often crucial to have information about one’s collaborators: successful driving requires understanding which side of the road to drive on. However, not every feature of collaborators is strategically relevant: the fine-grained acceleration of drivers may be ignored while maintaining optimal coordination. We show that there is a well-defined dichotomy between strategically relevant and irrelevant information. Moreover, we show that, in dynamic games, this dichotomy has a compact representation that can be efficiently computed via a Bellman backup operator. We apply this algorithm to analyze the strategically relevant information for tasks in both a standard and a partially observable version of the Overcooked environment. Theoretical and empirical results show that our algorithms are significantly more efficient than baselines. Videos are available at https://minknowledge.github.io.} }
Endnote
%0 Conference Paper %T Who Needs to Know? Minimal Knowledge for Optimal Coordination %A Niklas Lauffer %A Ameesh Shah %A Micah Carroll %A Michael D Dennis %A Stuart Russell %B Proceedings of the 40th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2023 %E Andreas Krause %E Emma Brunskill %E Kyunghyun Cho %E Barbara Engelhardt %E Sivan Sabato %E Jonathan Scarlett %F pmlr-v202-lauffer23a %I PMLR %P 18599--18613 %U https://proceedings.mlr.press/v202/lauffer23a.html %V 202 %X To optimally coordinate with others in cooperative games, it is often crucial to have information about one’s collaborators: successful driving requires understanding which side of the road to drive on. However, not every feature of collaborators is strategically relevant: the fine-grained acceleration of drivers may be ignored while maintaining optimal coordination. We show that there is a well-defined dichotomy between strategically relevant and irrelevant information. Moreover, we show that, in dynamic games, this dichotomy has a compact representation that can be efficiently computed via a Bellman backup operator. We apply this algorithm to analyze the strategically relevant information for tasks in both a standard and a partially observable version of the Overcooked environment. Theoretical and empirical results show that our algorithms are significantly more efficient than baselines. Videos are available at https://minknowledge.github.io.
APA
Lauffer, N., Shah, A., Carroll, M., Dennis, M.D. & Russell, S.. (2023). Who Needs to Know? Minimal Knowledge for Optimal Coordination. Proceedings of the 40th International Conference on Machine Learning, in Proceedings of Machine Learning Research 202:18599-18613 Available from https://proceedings.mlr.press/v202/lauffer23a.html.

Related Material