Reproduce “The Generalized Universal Law of Generalization” 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
Nick Chater, Paul Vitanyi
The authors derive an exponential relation between confusion probabilities and information distance, measured by the length of the shortest program transforming either item into the other. Under computability assumptions, it is universal up to additive constants among normalized, upper-semicomputable metrics. The result holds for almost all responses as lengths grow, not every pair, and is a derivation, not evidence that people use it.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.