Searcharxiv⌕ Search

arXiv subjects

J. G. Hoskins

Publications and source records attributed to J. G. Hoskins.

2 recordsLinked to original sources

Asymptotic Analysis of the Narrow Escape and Berg-Purcell problems on general three-dimensional domains with reactive boundary patches

We present an asymptotic analysis of two diffusive capture problems in general smooth closed three-dimensional geometries with multiple small reactive boundary patches of arbitrary shapes. (i) The narrow escape problem seeks to determine the escape rate of Brownian particles from an enclosed region through small boundary windows. (ii) The related Berg-Purcell (or narrow entrance) problem seeks to resolve the capture rate for signaling molecules diffusing outside the cell and entering through localized reactions at membrane-bound receptors. We obtain matched asymptotic solutions of these two problems and thus address the long-standing challenge of describing the role that curvature and local reactivities play in modulating diffusive capture rates. Our explicit expansions quantify local effects on diffusive capture through the sizes, shapes, and reactivities of the patches together with the principal curvatures of the manifold at each patch. In turn, we examine global effects on diffusive capture such as the spatial configuration of patches on the manifold, as encloded by the associated surface Neumann Green's function and its regular part. The accuracy of our asymptotic formulas is validated against a full numerical solution for an ellipsoidal domain. Overall, our results yield new insights on how geometry and stochasticity combine to shape the dynamics of various biological processes.

math.AP↗

Generative modeling via tensor train sketching

In this paper, we introduce a sketching algorithm for constructing a tensor train representation of a probability density from its samples. Our method deviates from the standard recursive SVD-based procedure for constructing a tensor train. Instead, we formulate and solve a sequence of small linear systems for the individual tensor train cores. This approach can avoid the curse of dimensionality that threatens both the algorithmic and sample complexities of the recovery problem. Specifically, for Markov models under natural conditions, we prove that the tensor cores can be recovered with a sample complexity that scales logarithmically in the dimensionality. Finally, we illustrate the performance of the method with several numerical experiments.

math.NA↗