Personalized Language Models via Privacy-Preserving Evolutionary Model Merging

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Kim, Kyuyoung, Shin, Jinwoo, Kim, Jaehyung
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914046617321472
author Kim, Kyuyoung
Shin, Jinwoo
Kim, Jaehyung
author_facet Kim, Kyuyoung
Shin, Jinwoo
Kim, Jaehyung
contents Personalization in language models aims to tailor model behavior to individual users or user groups. Prompt-based methods incorporate user preferences into queries, while training-based methods encode them into model parameters. Model merging has also been explored for personalization under limited data. However, existing methods often fail to directly optimize task-specific utility and lack explicit mechanisms for privacy preservation. To address the limitations, we propose Privacy-Preserving Model Merging via Evolutionary Algorithms (PriME), a novel personalization approach that employs gradient-free methods to directly optimize utility while reducing privacy risks. By integrating privacy preservation into the optimization objective, PriME creates personalized modules that effectively capture target user preferences while minimizing privacy risks for data-sharing users. Experiments on the LaMP benchmark show that PriME consistently outperforms a range of baselines, achieving up to a 45% improvement in task performance. Further analysis demonstrates that PriME achieves a superior privacy-utility trade-off compared to a prior state-of-the-art, with enhanced robustness to membership inference attacks and greater utility in capturing user preferences.
format Preprint
id arxiv_https___arxiv_org_abs_2503_18008
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Personalized Language Models via Privacy-Preserving Evolutionary Model Merging
Kim, Kyuyoung
Shin, Jinwoo
Kim, Jaehyung
Computation and Language
Neural and Evolutionary Computing
Personalization in language models aims to tailor model behavior to individual users or user groups. Prompt-based methods incorporate user preferences into queries, while training-based methods encode them into model parameters. Model merging has also been explored for personalization under limited data. However, existing methods often fail to directly optimize task-specific utility and lack explicit mechanisms for privacy preservation. To address the limitations, we propose Privacy-Preserving Model Merging via Evolutionary Algorithms (PriME), a novel personalization approach that employs gradient-free methods to directly optimize utility while reducing privacy risks. By integrating privacy preservation into the optimization objective, PriME creates personalized modules that effectively capture target user preferences while minimizing privacy risks for data-sharing users. Experiments on the LaMP benchmark show that PriME consistently outperforms a range of baselines, achieving up to a 45% improvement in task performance. Further analysis demonstrates that PriME achieves a superior privacy-utility trade-off compared to a prior state-of-the-art, with enhanced robustness to membership inference attacks and greater utility in capturing user preferences.
title Personalized Language Models via Privacy-Preserving Evolutionary Model Merging
topic Computation and Language
Neural and Evolutionary Computing
url https://arxiv.org/abs/2503.18008