Physics-informed distribution transformers via molecular dynamics and deep neural networks

Generating quasirandom points with high uniformity is a fundamental task in many fields. Existing number-theoretic approaches produce evenly distributed points in [0, 1]d in asymptotic sense but may not yield a good distribution for a given set size. It is also difficult to extend those techniques to other geometries like a disk or a manifold. In this paper, we present a novel physics-informed framework to transform a given set of points into a distribution with better uniformity. We model each point as a particle and assign the system with a potential energy. Upon minimizing the energy, the uniformity of distribution can be improved correspondingly. Two kinds of schemes are introduced: one based on molecular dynamics and another based on deep neural networks. The new physics-informed framework serves as a black-box transformer that is able to improve given distributions and can be easily extended to other geometries such as disks, spheres, complex manifolds, etc. Various experiments with different geometries are provided to demonstrate that the new framework is able to transform poorly distributed input into one with superior uniformity. (c) 2022 Elsevier Inc. All rights reserved.

Automated summary

Generating quasirandom points with high uniformity is a fundamental task in many fields. This paper presents a novel physics-informed framework that can transform a given set of points into a distribution with better uniformity. Two schemes based on molecular dynamics and deep neural networks are introduced. The new framework can be easily extended to other geometries and various experiments demonstrate its effectiveness.