MemPrivacy：面向邊雲代理的隱私保護個性化記憶管理

MemPrivacy: Privacy-Preserving Personalized Memory Management for Edge-Cloud Agents

May 10, 2026

作者: Yining Chen, Jihao Zhao, Bo Tang, Haofen Wang, Feiyu Xiong, Zhiyu Li

cs.AI

摘要

隨著基於大型語言模型的代理在邊緣-雲環境中日漸廣泛部署，個人化記憶已成為實現長期適應與以使用者為中心互動的關鍵技術。然而，雲端輔助的記憶管理會暴露敏感的個人資訊，而現有的隱私保護方法通常依賴於激進的遮罩處理，這種方式會移除與任務相關的語意，進而降低記憶的實用性與個人化品質。為了解決此問題，我們提出MemPrivacy方法，該方法在邊緣裝置上識別隱私敏感片段，將其替換為語意結構化的類型感知佔位符，供雲端進行記憶處理，並在需要時於本地端還原原始數值。透過將隱私保護與語意破壞加以解耦，MemPrivacy能在最小化敏感資料曝露的同時，保留記憶形成與檢索所需的有效資訊。我們也建構了用於系統性評估的MemPrivacy-Bench資料集，涵蓋200位使用者與超過52,000個隱私實例，並引入四層級隱私分類法以實現可配置的保護策略。實驗結果顯示，MemPrivacy在隱私資訊提取上展現出優異效能，大幅超越如GPT-5.2與Gemini-3.1-Pro等強大的通用模型，同時亦降低推論延遲。在多個廣泛使用的記憶系統中，MemPrivacy將實用性損失控制在1.6%以內，優於基準的遮罩策略。整體而言，MemPrivacy為邊緣-雲端代理提供了隱私保護與個人化記憶實用性之間的有效平衡，實現安全、實用且對使用者透明的部署方式。

English

As LLM-powered agents are increasingly deployed in edge-cloud environments, personalized memory has become a key enabler of long-term adaptation and user-centric interaction. However, cloud-assisted memory management exposes sensitive user information, while existing privacy protection methods typically rely on aggressive masking that removes task-relevant semantics and consequently degrades memory utility and personalization quality. To address this challenge, We propose MemPrivacy, which identifies privacy-sensitive spans on edge devices, replaces them with semantically structured type-aware placeholders for cloud-side memory processing, and restores the original values locally when needed. By decoupling privacy protection from semantic destruction, MemPrivacy minimizes sensitive data exposure while retaining the information required for effective memory formation and retrieval. We also construct MemPrivacy-Bench for systematic evaluation, a dataset covering 200 users and over 52k privacy instances, and introduce a four-level privacy taxonomy for configurable protection policies. Experiments show that MemPrivacy achieves strong performance in privacy information extraction, substantially surpassing strong general-purpose models such as GPT-5.2 and Gemini-3.1-Pro, while also reducing inference latency. Across multiple widely used memory systems, MemPrivacy limits utility loss to within 1.6%, outperforming baseline masking strategies. Overall, MemPrivacy offers an effective balance between privacy protection and personalized memory utility for edge-cloud agents, enabling secure, practical, and user-transparent deployment.