arXiv cs.AI· Songshi Liang, Ting-En Lin, Zixuan Huang, Yuchuan Wu, Yongbin Li, Zihe Wang, Rui Yan·· 1 小时前AI 评分28
MuEdit：一种面向跨域知识冲突的高效多任务编辑方法
MuEdit: An efficient multi-task editing method towards inter-domain knowledge conflicts
AI 导读
LLM 多任务知识编辑在顺序执行时会失效，论文将其归因于跨域零空间错位——异质任务产生不同的 Jacobian 几何结构，导致公共零空间急剧收缩。作者提出 Conflict Index 量化几何干扰，并提出 MuEdit，通过冲突感知排序和动态低秩近似扩展零空间来缓解多任务冲突。
正文 · 原文
Abstract:Large language models (LLMs) require frequent knowledge updates across heterogeneous domains, but existing editing methods fail when applied sequentially to multi-task settings. We trace this collapse to inter-domain null-space misalignment: heterogeneous tasks induce distinct Jacobian geometries, causing their common null space to shrink dramatically and leaving insufficient admissible directions for conflict-free updates. We propose the Conflict Index to quantify this geometric interference, and introduce Mu-Edit, which mitigates multi-task conflicts via (1) conflict-aware ordering to minimize cumulative interference, and (2) dynamic low-rank approximation to expand the null space when ordering alone is insufficient. Experiments across five functional domains and three backbones show that Mu-Edit significantly outperforms existing baselines in multi-task editing performance while preserving general capabilities, and remains effective under incremental task arrival.
| Comments: | 30 pages,accepted by NeurIPS 2026 |
| Subjects: | Artificial Intelligence (cs.AI); Computation and Language (cs.CL) |
| Cite as: | arXiv:2611.14910 [cs.AI] |
| (or arXiv:2611.14910v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2611.14910 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Liang Songshi [view email]
[v1]
Wed, 11 Nov 2026 02:06:19 UTC (2,646 KB)
来源：arXiv cs.AI · arxiv.org