{"id":"arxiv/2606.11183","url":"https://arxiv.org/abs/2606.11183","tags":["math.ST","math.DG","stat.ME"],"score":0,"title":"Nonparametric Riemannian Empirical Bayes, and Denoising Measurements on Manifolds","domain":"math.ST","source":"arxiv","authors":["Adam Quinn Jaffe","Leonardo V. Santoro","Bodhisattva Sen"],"reviews":0,"subject":"math.ST","summary":"We initiate the study of nonparametric empirical Bayes denoising methods in the setting where both the latent variables and their measurements lie on a compact Riemannian manifold, and where the likelihood is a Riemannian Gaussian distribution. Our starting point is a novel Tweedie-Eddington formula for Riemannian Gaussian mixture models which identifies a certain surrogate oracle denoiser in terms of the marginal distribution of the measurements; it avoids the explicit computation of the posterior Fréchet mean (as required by the Bayes denoiser) via a first-order approximation, hence we…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-06-09","reactions":{}}