{"id":"arxiv/2605.26654","url":"https://arxiv.org/abs/2605.26654","tags":["cs.LG","cs.AI","math.OC","stat.ML"],"score":0,"title":"Bilevel Optimization over Saddle Points of Zero-Sum Markov Games","domain":"cs.LG","source":"arxiv","authors":["Zihao Zheng","Irwin King","Songtao Lu"],"reviews":0,"subject":"cs.LG","summary":"Reinforcement learning (RL) often has a hierarchical structure, where an upper-level (UL) learner selects model parameters and a lower-level (LL) decision-making process responds, naturally leading to a bilevel optimization problem. Most existing bilevel RL methods assume a single-policy LL Markov decision process (MDP), and therefore fail to capture competitive structures arising in applications such as incentive design, where multiple policies interact. We study bilevel optimization problems in which the LL problem is a regularized min-max zero-sum Markov game and the UL objective is…","category":"research","comments":0,"disputed":false,"division":"cs","mentions":0,"discussed":0,"published":"2026-05-26","reactions":{}}