@inproceedings{chen-etal-2026-relaxing,
    title = "Relaxing the Constraints: A Dual-Importance Projection Mechanism for Lifelong Model Editing",
    author = "Chen, Zhenghai  and
      Xu, Senbin  and
      Tan, Jiaxi  and
      Wu, Xinhua  and
      Zhang, Yan  and
      Zheng, Xiawu  and
      Zhang, Shengchuan  and
      Li, Ke  and
      Zhao, Sicheng  and
      Cao, Liujuan  and
      Ji, Rongrong",
    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.1910/",
    pages = "38308--38333",
    ISBN = "979-8-89176-395-1",
    abstract = "Factual knowledge stored in Large Language Models (LLMs) inevitably becomes outdated or erroneous over time, making it critical to update these models without incurring the high cost of retraining. Existing sequential knowledge editing methods predominantly rely on strict orthogonal projection to preserve previously edited knowledge. However, this excessive constraint limits gradient expressiveness, resulting in a significant degradation of model generalization and overall performance as the number of edits increases. To address this challenge, we propose Dual-Importance Projection Editing (DipEdit). This method leverages Singular Value Decomposition (SVD) to identify critical gradient subspaces and introduces a dual mechanism comprising ``accumulated importance'' and ``projection importance.'' Unlike traditional approaches that enforce strict orthogonality, DipEdit dynamically scales gradient components parallel to key subspaces based on their projection importance rather than discarding them directly. This approach enhances the model{'}s adaptability to new knowledge while maximally preserving historical knowledge. Extensive experiments conducted on five mainstream LLMs using the ZsRE and Counterfact datasets demonstrate that DipEdit effectively handles thousands of sequential edits. The proposed method achieves an average comprehensive performance improvement of 10.36{\%} and effectively maintains the model{'}s general capabilities on downstream tasks. Code is available at: https://github.com/czhhhla/DipEdit."
}