Reproduce “On the Normalizing Constant of the Continuous Categorical Distribution” 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
Elliott Gordon-Rodriguez, Gabriel Loaiza-Ganem, Andres Potapczynski, John P. Cunningham
The authors study reliable computation of the continuous categorical distribution’s normalizing constant, which can lose precision when a closed-form sum cancels large terms. They recast the constant as an inverse Laplace transform and compare numerical methods against an arbitrary-precision oracle. De Hoog succeeds in some cases where direct evaluation fails, but no tested inversion method works beyond K=30.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.