@inproceedings{shen-etal-2026-metacognitive,
    title = "Metacognitive Self-Correction for Multi-Agent System via Prototype-Guided Next-Execution Reconstruction",
    author = "Shen, Xu  and
      Zhang, Qi  and
      Wang, Song  and
      Tan, Zhen  and
      Zhao, Xinyu  and
      Yao, Laura  and
      Tadiparthi, Vaishnav  and
      Mahjoub, Hossein Nourkhiz  and
      Pari, Ehsan Moradi  and
      Lee, Kwonjoon  and
      Chen, Tianlong",
    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.1168/",
    pages = "23320--23337",
    ISBN = "979-8-89176-395-1",
    abstract = "Large Language Model based multi-agent systems (MAS) excel at collaborative problem solving but remain brittle to cascading errors: a single faulty step can propagate across agents and disrupt the trajectory. In this paper, we present MASC, a metacognitive framework that endows MAS with real-time, unsupervised, step-level error detection and self-correction. MASC rethinks detection as history-conditioned anomaly scoring via two complementary designs: (1) Next-Execution Reconstruction, which predicts the embedding of the next step from the query and interaction history to capture causal consistency, and (2) Prototype-Guided Enhancement, which learns a prototype prior over normal-step embeddings and uses it to stabilize reconstruction and anomaly scoring under sparse context (e.g., early steps). When an anomaly step is flagged, MASC triggers a correction agent to revise the acting agent{'}s output before information flows downstream. On the Who When benchmark, MASC consistently outperforms all baselines, achieving up to 7.8{\%} AUC-ROC improvement in the challenging w/o GT setting, and further delivers consistent gains on AgentErrorBench. When plugged into diverse MAS frameworks, it delivers consistent end-to-end gains across architectures, confirming that our metacognitive monitoring and targeted correction can mitigate error propagation with minimal overhead."
}