Reproduce “Learning What to Retain: Gated-Memory Routing for Efficient Collaboration in Multi-Agent 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
Rakibul Hasan Rajib, Mengxin Zheng, Qian Lou
Gated-Memory Routing coordinates specialist language-model agents by selecting roles and backbones, retrieving relevant records, writing nonredundant responses to execution memory, and halting adaptively. Across five closed-domain benchmarks, it averages 77.73% test accuracy versus 75.29% for Puppeteer with Qwen-2.5-32B. It leads on four tasks, but not MBPP, where MacNet-Chain scores 82.66% versus 79.60%.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.