@inproceedings{qian-etal-2026-improved,
    title = "Improved Policy Optimization for Mixture-of-Experts Models: Importance Sampling and Rewarding from an Expert-Centric Perspective",
    author = "Qian, Yining  and
      Li, Jinpeng  and
      Mi, Fei  and
      Shang, Lifeng  and
      Zhang, Xiang",
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {ACL} 2026",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl/2026.findings-acl.1944/",
    pages = "39029--39038",
    ISBN = "979-8-89176-395-1",
    abstract = "Reinforcement learning (RL) has demonstrated considerable promise in enhancing large language models. However, its application to Mixture-of-Experts (MoE) architectures is frequently hindered by training instability, primarily stemming from token-level misalignment in expert assignments between current and behavior policies. Existing approaches often oscillate between overly coarse sequence-level importance sampling, which ignores token-specific discrepancies, and restrictive expert-selection constraints that suppress beneficial policy exploration. To bridge this gap, we propose Expert Relative Policy Optimization (ERPO), which introduces expert-level importance sampling. By grouping tokens according to their routing assignments, ERPO mitigates the high variance of token-level importance sampling while overcoming the token-agnostic limitations of sequence-level methods. Furthermore, ERPO leverages this expert-centric granularity to introduce an Expert-Selection Entropy Reward, which dynamically adjusts routing uncertainty based on task-specific feedback. This unique mechanism ensures a rigorous alignment between reward signals and policy updates{---}a capability inherently unattainable by traditional importance sampling methods. Experimental results demonstrate that ERPO significantly outperforms strong baselines across multiple reasoning tasks, highlighting the efficacy of tailoring RL objectives to the structural inductive biases of MoE models."
}