@inproceedings{ai-etal-2026-cognitive,
    title = "Cognitive Scaffold: From Fluid Context to Crystallized Memory for Long-Horizon {D}eep{R}esearch Agents",
    author = "Ai, Qiuyuan  and
      Fu, Zenghuang  and
      Li, Zhaoyang  and
      Jiang, Ping  and
      Wu, Haoyu  and
      Song, Jie  and
      He, Guannan",
    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.1170/",
    pages = "25526--25542",
    ISBN = "979-8-89176-390-6",
    abstract = "Scaling LLM-based agents to long-horizon deep research is constrained by the context-noise trade-off, where linear history accumulation degrades reasoning and dilutes fine-grained evidence. To address this, we introduce the Cognitive Scaffold, a factorized memory architecture that decouples the cognitive state into a Fluid Working Context for immediate reasoning and a persistent Knowledge Graph for long-term retention. Unlike unstructured summarization, our framework employs a Rejection Sampling Fine-Tuning (RFT) pipeline to crystallize saturated context into structured event snapshots, strictly enforcing atomic constraints to preserve numerical values and entities. During reasoning, a thought-driven dual-path retrieval mechanism enables the agent to proactively recover precise evidence. Empirical evaluations on Xbench-DeepSearch, BrowseComp-ZH, and GAIA demonstrate that Cognitive Scaffold consistently outperforms baselines, achieving 74.7{\%} Avg@3 and 87.0{\%} Pass@3 on Xbench-DeepSearch, 48.5{\%} Avg@3 and 65.9{\%} Pass@3 on BrowseComp-ZH, and 72.8{\%} Avg@3 and 88.3{\%} Pass@3 on GAIA, while reducing compression hallucinations to 5.3{\%}. We open-source our codebase to facilitate future research."
}