本テーマでは、ゲームの均衡を計算する学習アルゴリズムを研究する。特に、更新される戦略そのものがナッシュ均衡へ収束するようなアルゴリズムの構築を目指す。その一環として、利得摂動を用いて学習ダイナミクスを安定化する方法を扱う。
研究の焦点#
- 学習アルゴリズムはどのようにしてナッシュ均衡へ高速に収束できるか
- 利得摂動を用いて学習ダイナミクスをどのように安定化できるか
- マルチエージェントの学習ダイナミクスはいつ収束しないのか
Learning Dynamics and Equilibrium Computation in Games
ナッシュ均衡へ高速に収束する学習アルゴリズムとは？
関連論文
ICML 2026
Asymmetric Perturbation in Solving Bilinear Saddle-Point Optimization
テーマ:Learning in GamesNeurIPS 2025
Last Iterate Convergence in Monotone Mean Field Games
テーマ:Learning in GamesNeurIPS 2025
Learning from Delayed Feedback in Games via Extra Prediction
テーマ:Learning in GamesAAMAS 2025 (Full paper)
Global Behavior of Learning Dynamics in Zero-Sum Games with Memory Asymmetry
テーマ:Learning in GamesAAMAS 2025 (Extended abstract)
Nash Equilibrium and Learning Dynamics in Three-Player Matching m-Action Games
テーマ:Learning in GamesICLR 2025
Boosting Perturbed Gradient Ascent for Last-Iterate Convergence in Games
テーマ:Learning in GamesAAAI 2025
Approximate State Abstraction for Markov Games
テーマ:Learning in GamesAAAI 2025
Synchronization behind Learning in Periodic Zero-Sum Games Triggers Divergence from Nash equilibrium
テーマ:Learning in GamesarXiv
The Power of Perturbation under Sampling in Solving Extensive-Form Games
テーマ:Learning in GamesICML 2024
Adaptively Perturbed Mirror Descent for Learning in Games
テーマ:Learning in GamesAAAI 2024
Memory Asymmetry Creates Heteroclinic Orbits to Nash Equilibrium in Learning in Zero-Sum Games
テーマ:Learning in GamesIJCAI 2023
Learning in Multi-Memory Games Triggers Complex Dynamics Diverging from Nash Equilibrium
テーマ:Learning in GamesAISTATS 2023
Last-Iterate Convergence with Full and Noisy Feedback in Two-Player Zero-Sum Games
テーマ:Learning in GamesUAI 2022
Mutation-Driven Follow the Regularized Leader for Last-Iterate Convergence in Zero-Sum Games
テーマ:Learning in GamesAAAI 2022 Workshop on Reinforcement Learning in Games (Oral Presentation)
Computing Strategies of American Football via Counterfactual Regret Minimization
テーマ:Learning in GamesAAMAS 2021 (Full Paper)