{"id":"arxiv/2604.13563","url":"https://arxiv.org/abs/2604.13563","tags":["math.NA","math.ST"],"score":0,"title":"Covariance-Informed Subspace: an Adaptive Gradient-Free Input Dimension Reduction Method for Bayesian Inference","domain":"math.NA","source":"arxiv","authors":["Nadège Polette","Olivier Le Maître","Pierre Sochala","Alexandrine Gesret"],"reviews":0,"subject":"math.NA","summary":"This paper addresses the challenge of dimension reduction (DR) in Bayesian inference of high-resolution two-or three-dimensional fields, where a priori parametrizations require a large number of terms. The underlying idea is common to state-of-the-art methods in which the parameter space is decomposed into two subspaces, one informed by the likelihood and one constrained by the prior. DR techniques generally use gradient information from the log-likelihood to derive the corresponding subspaces. However, the gradient may be unavailable or expensive to compute accurately, for instance in the…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-04-15","reactions":{}}