Reproduce “A latent dimension of Condorcet's jury theorem for multiple AI advisers” 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
Kazutoshi Sasahara, Aoi Naito, Ryo Fujie
The authors model independent binary-question panels with equal adviser accuracy above chance. They compare majority-vote accuracy with visible dissent, defined as at least two advisers on each side. At accuracy 0.9, a panel of 20 has majority accuracy above 99.999%, yet visible dissent occurs about 61% of the time. These are model results, not evidence about deployed systems.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.