Personalized Language Models via Privacy-Preserving Evolutionary Model Merging

• URL: http://arxiv.org/abs/2503.18008v2
• Date: Fri, 19 Sep 2025 07:12:43 GMT
• Title: Personalized Language Models via Privacy-Preserving Evolutionary Model Merging
• Authors: Kyuyoung Kim, Jinwoo Shin, Jaehyung Kim,
• Abstract summary: Personalization in language models aims to tailor model behavior to individual users or user groups.<n>We propose Privacy-Preserving Model Merging via Evolutionary Algorithms (PriME)<n>PriME employs gradient-free methods to directly optimize utility while reducing privacy risks.<n>Experiments on the LaMP benchmark show that PriME consistently outperforms a range of baselines, achieving up to a 45% improvement in task performance.
• Score: 53.97323896430374
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: 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.

Related papers

• Synthetic Interaction Data for Scalable Personalization in Large Language Models [67.31884245564086]
  We introduce a high-fidelity synthetic data generation framework called PersonaGym.<n>Unlike prior work that treats personalization as static persona-preference pairs, PersonaGym models a dynamic preference process.<n>We release PersonaAtlas, a large-scale, high-quality, and diverse synthetic dataset of high-fidelity multi-turn personalized interaction trajectories.
  arXiv  Detail & Related papers  (2026-02-12T20:41:22Z)
• One Adapts to Any: Meta Reward Modeling for Personalized LLM Alignment [55.86333374784959]
  We argue that addressing these constraints requires a paradigm shift from fitting data to learn user preferences to learn the process of preference adaptation.<n>We propose Meta Reward Modeling (MRM), which reformulates personalized reward modeling as a meta-learning problem.<n>We show that MRM enhances few-shot personalization, improves user robustness, and consistently outperforms baselines.
  arXiv  Detail & Related papers  (2026-01-26T17:55:52Z)
• PRISP: Privacy-Safe Few-Shot Personalization via Lightweight Adaptation [21.467360472787593]
  PRISP is a lightweight and privacy-safe personalization framework.<n>It exploits a Text-to-LoRA hypernetwork to generate task-aware LoRA parameters from task descriptions.<n>Experiments on a few-shot variant of the LaMP benchmark demonstrate that PRISP achieves strong overall performance.
  arXiv  Detail & Related papers  (2026-01-10T07:34:28Z)
• PersonaAgent: When Large Language Model Agents Meet Personalization at Test Time [87.99027488664282]
  PersonaAgent is a framework designed to address versatile personalization tasks.<n>It integrates a personalized memory module and a personalized action module.<n>Test-time user-preference alignment strategy ensures real-time user preference alignment.
  arXiv  Detail & Related papers  (2025-06-06T17:29:49Z)
• NextQuill: Causal Preference Modeling for Enhancing LLM Personalization [82.15961484963256]
  We introduce NextQuill, a novel personalization framework grounded in causal preference modeling.<n>Building on this insight, NextQuill introduces two complementary alignment strategies.<n> Experiments across multiple personalization benchmarks demonstrate that NextQuill significantly improves personalization quality.
  arXiv  Detail & Related papers  (2025-06-03T02:08:55Z)
• Improved Algorithms for Differentially Private Language Model Alignment [12.611907955831597]
  We propose novel algorithms for privacy-preserving alignment.<n>Our framework can be deployed on two celebrated alignment techniques.<n>One of our algorithms, DP-AdamW, combined with DPO, surpasses existing methods, improving alignment quality by up to 15% under moderate privacy budgets.
  arXiv  Detail & Related papers  (2025-05-13T16:18:59Z)
• User Behavior Analysis in Privacy Protection with Large Language Models: A Study on Privacy Preferences with Limited Data [5.152440245370642]
  This study explores how large language models (LLMs) can analyze user behavior related to privacy protection in scenarios with limited data.<n>The research utilizes anonymized user privacy settings data, survey responses, and simulated data.<n> Experimental results demonstrate that, even with limited data, LLMs significantly improve the accuracy of privacy preference modeling.
  arXiv  Detail & Related papers  (2025-05-08T04:42:17Z)
• Adaptive Clipping for Privacy-Preserving Few-Shot Learning: Enhancing Generalization with Limited Data [12.614480013684759]
  We introduce a novel approach called Meta-Clip to enhance the utility of privacy-preserving few-shot learning methods. By dynamically adjusting clipping thresholds during the training process, our Adaptive Clipping method provides fine-grained control over the disclosure of sensitive information. We demonstrate the effectiveness of our approach in minimizing utility degradation, showcasing a superior privacy-preserving trade-off compared to existing privacy-preserving techniques.
  arXiv  Detail & Related papers  (2025-03-27T05:14:18Z)
• Multi-Objective Optimization for Privacy-Utility Balance in Differentially Private Federated Learning [12.278668095136098]
  Federated learning (FL) enables collaborative model training across distributed clients without sharing raw data. We propose an adaptive clipping mechanism that dynamically adjusts the clipping norm using a multi-objective optimization framework. Our results show that adaptive clipping consistently outperforms fixed-clipping baselines, achieving improved accuracy under the same privacy constraints.
  arXiv  Detail & Related papers  (2025-03-27T04:57:05Z)
• Measuring What Makes You Unique: Difference-Aware User Modeling for Enhancing LLM Personalization [68.79814761867314]
  We propose Difference-aware Personalization Learning (DPL) to enhance Large Language Models (LLMs) personalization.<n>DPL strategically selects representative users for comparison and establishes a structured standard to extract task-relevant differences.<n>Experiments on real-world datasets demonstrate that DPL significantly enhances LLM personalization.
  arXiv  Detail & Related papers  (2025-03-04T09:53:26Z)
• Advancing Personalized Federated Learning: Integrative Approaches with AI for Enhanced Privacy and Customization [0.0]
  This paper proposes a novel approach that enhances PFL with cutting-edge AI techniques.<n>We present a model that boosts the performance of individual client models and ensures robust privacy-preserving mechanisms.<n>This work paves the way for a new era of truly personalized and privacy-conscious AI systems.
  arXiv  Detail & Related papers  (2025-01-30T07:03:29Z)
• Few-shot Steerable Alignment: Adapting Rewards and LLM Policies with Neural Processes [50.544186914115045]
  Large language models (LLMs) are increasingly embedded in everyday applications. Ensuring their alignment with the diverse preferences of individual users has become a critical challenge. We present a novel framework for few-shot steerable alignment.
  arXiv  Detail & Related papers  (2024-12-18T16:14:59Z)
• Efficient and Private: Memorisation under differentially private parameter-efficient fine-tuning in language models [2.3281513013731145]
  Fine-tuning large language models (LLMs) for specific tasks introduces privacy risks, as models may inadvertently memorise and leak sensitive training data. Differential Privacy (DP) offers a solution to mitigate these risks, but introduces significant computational and performance trade-offs. We show that PEFT methods achieve comparable performance to standard fine-tuning while requiring fewer parameters and significantly reducing privacy leakage.
  arXiv  Detail & Related papers  (2024-11-24T13:17:36Z)
• Guided Profile Generation Improves Personalization with LLMs [3.2685922749445617]
  In modern commercial systems, including Recommendation, Ranking, and E-Commerce platforms, there is a trend towards incorporating Personalization context as input into Large Language Models (LLMs) We propose Guided Profile Generation (GPG), a general method designed to generate personal profiles in natural language. Our experimental results show that GPG improves LLM's personalization ability across different tasks, for example, it increases 37% accuracy in predicting personal preference compared to directly feeding the LLMs with raw personal context.
  arXiv  Detail & Related papers  (2024-09-19T21:29:56Z)
• Exposing Privacy Gaps: Membership Inference Attack on Preference Data for LLM Alignment [8.028743532294532]
  We introduce a novel reference-based attack framework specifically for analyzing preference data called PREMIA. We provide empirical evidence that DPO models are more vulnerable to MIA compared to PPO models.
  arXiv  Detail & Related papers  (2024-07-08T22:53:23Z)
• Mind the Privacy Unit! User-Level Differential Privacy for Language Model Fine-Tuning [62.224804688233]
  differential privacy (DP) offers a promising solution by ensuring models are 'almost indistinguishable' with or without any particular privacy unit. We study user-level DP motivated by applications where it necessary to ensure uniform privacy protection across users.
  arXiv  Detail & Related papers  (2024-06-20T13:54:32Z)
• Differentially Private Fine-Tuning of Diffusion Models [22.454127503937883]
  The integration of Differential Privacy with diffusion models (DMs) presents a promising yet challenging frontier. Recent developments in this field have highlighted the potential for generating high-quality synthetic data by pre-training on public data. We propose a strategy optimized for private diffusion models, which minimizes the number of trainable parameters to enhance the privacy-utility trade-off.
  arXiv  Detail & Related papers  (2024-06-03T14:18:04Z)
• The Data Minimization Principle in Machine Learning [61.17813282782266]
  Data minimization aims to reduce the amount of data collected, processed or retained. It has been endorsed by various global data protection regulations. However, its practical implementation remains a challenge due to the lack of a rigorous formulation.
  arXiv  Detail & Related papers  (2024-05-29T19:40:27Z)
• Differentially Private Deep Model-Based Reinforcement Learning [47.651861502104715]
  We introduce PriMORL, a model-based RL algorithm with formal differential privacy guarantees. PriMORL learns an ensemble of trajectory-level DP models of the environment from offline data.
  arXiv  Detail & Related papers  (2024-02-08T10:05:11Z)
• Private Set Generation with Discriminative Information [63.851085173614]
  Differentially private data generation is a promising solution to the data privacy challenge. Existing private generative models are struggling with the utility of synthetic samples. We introduce a simple yet effective method that greatly improves the sample utility of state-of-the-art approaches.
  arXiv  Detail & Related papers  (2022-11-07T10:02:55Z)
• Group privacy for personalized federated learning [4.30484058393522]
  Federated learning is a type of collaborative machine learning, where participating clients process their data locally, sharing only updates to the collaborative model. We propose a method to provide group privacy guarantees exploiting some key properties of $d$-privacy.
  arXiv  Detail & Related papers  (2022-06-07T15:43:45Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.