Reproduce “MADE: Belief-Driven Dual-Agent Coordination for Autonomous Model Deployment” 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
Yicheng Liu, Bolin Zhang, Weiran Liu, Yakun Zhang, Yangqin Jiang, Zhiying Tu, Dianhui Chu
MADE turns research-model repositories and artifacts into callable APIs by coordinating deployment phases and revisiting earlier choices when runtime evidence indicates a fault. On M2ABench, it successfully deploys 68.85% of 122 model resources, exceeding SWE-agent by 13.93 percentage points. Success requires structurally matching endpoint outputs on all five test cases; this measures deployment, not predictive accuracy.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.