Learning from Others, Acting for You: Cross-User Memory Sharing for LLM Agents

2026-09-26

AI Summary

This study addresses the limitations of experience reuse among LLM agents and the user preference conflicts arising from shared memory by proposing the ShareMem architecture. Its core innovation lies in decoupling "action guidance" from "preference values," thereby enabling secure cross-user experience transfer. Furthermore, ShareMem incorporates a two-stage memory consolidation mechanism, scope-prioritized retrieval, and user-bound channels to ensure that task execution strictly adheres to the recipient's local preferences. Experimental evaluations across web navigation, personalized interaction, and coding benchmarks demonstrate that the proposed approach significantly improves step success rates, task completion rates, and code quality.

Abstract

Large language model (LLM) agents serving different users often solve related tasks, yet separate user histories can leave reusable experience inaccessible to other agents. Pooling memories expands access but risks transferring preferences that conflict with the receiving user's requirements. We introduce ShareMem, a memory architecture that shares reusable experience while grounding its application in the receiving user's own preferences. Shared experiences indicate how to act and which preferences to consult; the receiving user's memory supplies their concrete values. Two-stage consolidation refines experience locally before integrating accepted edits into a shared pool. During execution, scope-first retrieval jointly selects local and shared experiences under a common entry budget, while a user-bound channel supports initial and agent-initiated preference retrieval. We evaluate ShareMem across web navigation (Mind2Web), online personalized interaction (VitaBench~2.0), and multi-session coding (MemoryCode) with four backbone models. It improves step success, average task success, and dialogue-macro coding scores, respectively, over matched user-local memory across all four models. Ablations favor two-stage consolidation for smaller shared pools, lower induction token usage, and better downstream performance, and support complementarity between experience guidance and active preference retrieval. Further analyses show that sharing helps most when relevant local experience is scarce, while source quality and cross-user preference interference limit useful transfer.

Problem

• LLM agents
• cross-user memory sharing
• preference conflict
• reusable experience
• user personalization

Innovation

• Cross-User Memory Sharing
• ShareMem Architecture
• Two-Stage Consolidation
• Scope-First Retrieval
• Preference Grounding