Reproduce “Scientist World: Grow a Research World for AI-Scientists” 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
Taiming Lu, Sachin Konan, Zhuang Liu
Scientist World turns empirical AI research papers into runnable environments where agents reproduce methods, set improvement targets, package tasks, and verify them. Its release contains 1,000 environments, while tests cover ten lower-compute environments. Attempts improved over reproduced methods in 84% of Improve runs and 74% of Innovate runs, measured by held-out gains; Innovate has more regressions.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.