Reproduce “A New Bound on the Cumulant Generating Function of Dirichlet Processes” 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
Pierre Perrault, Denis Belomestny, Pierre Ménard, Éric Moulines, Alexey Naumov, Daniil Tiapkin, Michal Valko
The authors bound the cumulant-generating function of an integral against a Dirichlet process for sums of independent random probability measures. They prove superadditivity and combine it with a large-deviation limit to obtain, for continuous functions on a compact space and every positive concentration parameter, a reverse-KL variational upper bound. It is asymptotically optimal as the parameter grows; this is a theoretical result.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.