EMSEdit: Efficient Multi-Step Meta-Learning-based Model Editing

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Li, Xiaopeng, Li, Shasha, Wang, Xi, Song, Shezheng, Ji, Bin, Wang, Shangwen, Ma, Jun, Liu, Xiaodong, Liu, Mina, Yu, Jie
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914288274243584
author Li, Xiaopeng
Li, Shasha
Wang, Xi
Song, Shezheng
Ji, Bin
Wang, Shangwen
Ma, Jun
Liu, Xiaodong
Liu, Mina
Yu, Jie
author_facet Li, Xiaopeng
Li, Shasha
Wang, Xi
Song, Shezheng
Ji, Bin
Wang, Shangwen
Ma, Jun
Liu, Xiaodong
Liu, Mina
Yu, Jie
contents Large Language Models (LLMs) power numerous AI applications, yet updating their knowledge remains costly. Model editing provides a lightweight alternative through targeted parameter modifications, with meta-learning-based model editing (MLME) demonstrating strong effectiveness and efficiency. However, we find that MLME struggles in low-data regimes and incurs high training costs due to the use of KL divergence. To address these issues, we propose $\textbf{E}$fficient $\textbf{M}$ulti-$\textbf{S}$tep $\textbf{Edit (EMSEdit)}$, which leverages multi-step backpropagation (MSBP) to effectively capture gradient-activation mapping patterns within editing samples, performs multi-step edits per sample to enhance editing performance under limited data, and introduces norm-based regularization to preserve unedited knowledge while improving training efficiency. Experiments on two datasets and three LLMs show that EMSEdit consistently outperforms state-of-the-art methods in both sequential and batch editing. Moreover, MSBP can be seamlessly integrated into existing approaches to yield additional performance gains. Further experiments on a multi-hop reasoning editing task demonstrate EMSEdit's robustness in handling complex edits, while ablation studies validate the contribution of each design component. Our code is available at https://github.com/xpq-tech/emsedit.
format Preprint
id arxiv_https___arxiv_org_abs_2508_04012
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle EMSEdit: Efficient Multi-Step Meta-Learning-based Model Editing
Li, Xiaopeng
Li, Shasha
Wang, Xi
Song, Shezheng
Ji, Bin
Wang, Shangwen
Ma, Jun
Liu, Xiaodong
Liu, Mina
Yu, Jie
Computation and Language
Artificial Intelligence
Machine Learning
Large Language Models (LLMs) power numerous AI applications, yet updating their knowledge remains costly. Model editing provides a lightweight alternative through targeted parameter modifications, with meta-learning-based model editing (MLME) demonstrating strong effectiveness and efficiency. However, we find that MLME struggles in low-data regimes and incurs high training costs due to the use of KL divergence. To address these issues, we propose $\textbf{E}$fficient $\textbf{M}$ulti-$\textbf{S}$tep $\textbf{Edit (EMSEdit)}$, which leverages multi-step backpropagation (MSBP) to effectively capture gradient-activation mapping patterns within editing samples, performs multi-step edits per sample to enhance editing performance under limited data, and introduces norm-based regularization to preserve unedited knowledge while improving training efficiency. Experiments on two datasets and three LLMs show that EMSEdit consistently outperforms state-of-the-art methods in both sequential and batch editing. Moreover, MSBP can be seamlessly integrated into existing approaches to yield additional performance gains. Further experiments on a multi-hop reasoning editing task demonstrate EMSEdit's robustness in handling complex edits, while ablation studies validate the contribution of each design component. Our code is available at https://github.com/xpq-tech/emsedit.
title EMSEdit: Efficient Multi-Step Meta-Learning-based Model Editing
topic Computation and Language
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2508.04012