Reproduce “Self-Driving Datasets: From 20 Million Papers to Nuanced Biomedical Knowledge at Scale” 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
Haydn Jones, Yimeng Zeng, Alden Rose, Li S. Yifei, Yining Huang, Kaiwen Wu, Jiaming Liang, Maggie Ziyu Huan and 6 more
Starling builds searchable biomedical datasets by retrieving papers and extracting records with experimental context and supporting passages. Its system uses entity tagging and iterative retrieval, then filters records with a judge. Across six tasks it reports about 6.3 million records. In a selected human audit, confirmed-wrong rates were 16.5% for curated BBB_Martins labels and 1.7% for Starling, both lower bounds.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.