@inproceedings{xu-etal-2026-learning,
    title = "Learning How and What to Memorize: Cognition-Inspired Two-Stage Optimization for Evolving Memory",
    author = "Xu, Derong  and
      Liu, Shuochen  and
      Luo, Pengfei  and
      Jia, Pengyue  and
      Zhang, Yingyi  and
      Wen, Yi  and
      Deng, Yimin  and
      Zhang, Wenlin  and
      Chen, Enhong  and
      Zhao, Xiangyu  and
      Xu, Tong",
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Proceedings of the 64th Annual Meeting of the {A}ssociation for {C}omputational {L}inguistics (Volume 1: Long Papers)",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl/2026.acl-long.2084/",
    pages = "44987--45011",
    ISBN = "979-8-89176-390-6",
    abstract = "Large language model (LLM) agents require long-term user memory for consistent personalization, but limited context windows hinder tracking evolving preferences over long interactions. Existing memory systems mainly rely on static, hand-crafted update rules; although reinforcement learning (RL)-based agents learn memory updates, sparse outcome rewards provide weak supervision, resulting in unstable long-horizon optimization. Drawing on memory schema theory and the functional division between prefrontal regions and hippocampus regions, we introduce MemCoE, a cognition-inspired two-stage optimization framework that learns how memory should be organized and what information to update. In the first stage, we propose Memory Guideline Induction to optimize a global guideline via contrastive feedback interpreted as textual gradients; in the second stage, Guideline-Aligned Memory Policy Optimization uses the induced guideline to define structured process rewards and performs multi-turn RL to learn a guideline-following memory evolution policy. We evaluate on three personalization memory benchmarks, covering explicit and implicit preferences as well as different sizes and noise levels, and observe consistent improvements over strong baselines with favorable robustness, transferability, and efficiency[https://github.com/Applied-Machine-Learning-Lab/ACL2026{\_}MemCoE]."
}