@inproceedings{zheng-etal-2026-dkme,
    title = "{DKME}: Rethinking Coupled Knowledge Memory for Lifelong Model Editing of Large Language Models",
    author = "Zheng, Guanyu  and
      Zhenyu, Wang  and
      Tingting, He  and
      Wang, Xv  and
      Wang, Haochang  and
      Huang, Yaokai  and
      Zhao, Tiejun",
    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.792/",
    pages = "16128--16150",
    ISBN = "979-8-89176-395-1",
    abstract = "Lifelong knowledge editing aims to inject a stream of factual updates into large language models (LLMs) without retraining, yet existing memory-based editors often suffer from catastrophic forgetting as edits accumulate. We argue that a key factor is the coupled knowledge memory mechanism, where addressing (routing) and storage (writing via memory-module updates) are entangled. This entanglement makes it difficult to confine the effects of each edit to its intended scope, particularly in multi-domain and associated-fact editing streams, where updates either span diverse semantic domains or repeatedly modify related attributes of the same subject. Consequently, updating memory for one edit inadvertently alters the routing and stored representations of previously injected edits, leading to catastrophic forgetting as edits accumulate. We propose **DKME**, which decouples addressing from storage via two stages: decoupled semantic addressing learns a fact-aware manifold for scope-aware routing, and partitioned memory storage localizes edits to memory partitions identified by unsupervised clustering in the embedding space. Experiments on three benchmarks, including HalluEditBench, CKnowEdit, and WikiData$_\text{counterfact}$, demonstrate that DKME consistently achieves a more favorable trade-off between editing success and locality compared to baselines, while maintaining more stable performance as the edit scale increases."
}