Reproduce “AIM: A Privacy-Aware Interoperable Memory Framework for Multi-Agent Multi-User LLM Systems” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Zachary Johnson, Nigel Boachie Kumankumah, Somya Chatterjee, Tejas Sathyamurthi, Min Chen, Xinyi Alice Li, Xiao Wang, Emily Morgan Gelchie and 3 more
AIM separates visibility classification from access control in a memory service. An LLM labels memories private or public; the index limits private memories to owners and enforces write ownership. On synthetic MUMBench, GPT-5.4 reached 96.0% visibility accuracy, while GPT-5.4-mini reached 58.8% strict and 70.5% state-aware operation accuracy. Retrieval relevance peaked at 45.0%. Neither privacy nor answer quality was evaluated.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.