eprintid: 10187244 rev_number: 12 eprint_status: archive userid: 699 dir: disk0/10/18/72/44 datestamp: 2024-02-15 13:55:55 lastmod: 2024-02-15 13:55:55 status_changed: 2024-02-15 13:55:55 type: proceedings_section metadata_visibility: show sword_depositor: 699 creators_name: Marris, Luke creators_name: Muller, Paul creators_name: Lanctot, Marc creators_name: Tuyls, Karl creators_name: Graepel, Thore title: Multi-Agent Training beyond Zero-Sum with Correlated Equilibrium Meta-Solvers ispublished: pub divisions: UCL divisions: B04 divisions: C05 divisions: F48 note: © The Authors 2023. Original content in this paper is licensed under the terms of the Creative Commons Attribution 4.0 International (CC BY 4.0) Licence (https://creativecommons.org/licenses/by/4.0/). abstract: Two-player, constant-sum games are well studied in the literature, but there has been limited progress outside of this setting. We propose Joint Policy-Space Response Oracles (JPSRO), an algorithm for training agents in n-player, general-sum extensive form games, which provably converges to an equilibrium. We further suggest correlated equilibria (CE) as promising meta-solvers, and propose a novel solution concept Maximum Gini Correlated Equilibrium (MGCE), a principled and computationally efficient family of solutions for solving the correlated equilibrium selection problem. We conduct several experiments using CE meta-solvers for JPSRO and demonstrate convergence on n-player, general-sum games. date: 2021 date_type: published publisher: Proceedings of Machine Learning Research official_url: https://proceedings.mlr.press/v139/marris21a.html oa_status: green full_text_type: pub language: eng primo: open primo_central: open_green verified: verified_manual elements_id: 1876674 lyricists_name: Graepel, Thore lyricists_name: Marris, Luke lyricists_id: TGRAE40 lyricists_id: LMARR01 actors_name: Flynn, Bernadette actors_id: BFFLY94 actors_role: owner full_text_status: public publication: Proceedings of the 38th International Conference on Machine Learning volume: 139 pagerange: 7480-7491 pages: 12 event_title: The 38th International Conference on Machine Learning book_title: Proceedings of the 38th International Conference on Machine Learning editors_name: Meila, M editors_name: Zhang, T citation: Marris, Luke; Muller, Paul; Lanctot, Marc; Tuyls, Karl; Graepel, Thore; (2021) Multi-Agent Training beyond Zero-Sum with Correlated Equilibrium Meta-Solvers. In: Meila, M and Zhang, T, (eds.) Proceedings of the 38th International Conference on Machine Learning. (pp. pp. 7480-7491). Proceedings of Machine Learning Research Green open access document_url: https://discovery.ucl.ac.uk/id/eprint/10187244/1/marris21a.pdf