{"id":"arxiv/2606.04190","url":"https://arxiv.org/abs/2606.04190","tags":["math.NA","math.FA"],"score":0,"title":"Sampling and reconstruction of convex functions","domain":"math.NA","source":"arxiv","authors":["Andrea Bonito","Albert Cohen","Wolfgang Dahmen","Ronald Devore","Guergana Petrova","Jonathan W. Siegel"],"reviews":0,"subject":"math.NA","summary":"We discuss optimal recovery for classes of multivariate convex functions from given point samples, as well as the sampling numbers of these classes, corresponding to optimal sample choices. Upper and lower bounds for either variant are established when the reconstruction error is measured in $L_p$ for $1\\leq p\\leq \\infty$. These bounds match, sometimes up to logarithmic factors, and therefore characterize the respective optimal rate of decay. For classical smoothness classes such as Sobolev, Hölder or Besov spaces, it is well known that the optimal decay rate of sampling numbers can be…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-06-02","reactions":{}}