BZPEERPapers, people, evidence and research pathsDiscover
Research tools
Network
Opportunities
Account
Source author record
Scott T. Miller appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.
Catalog footprint
Research graph
Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.
BZPEER is loading the nearby papers, people, topics and institutions for this page.
Published work
We detail how incorporating physics into neural network design can significantly improve the learning and forecasting of dynamical systems, even nonlinear systems of many dimensions. A map building perspective elucidates the superiority of Hamiltonian neural networks over conventional neural networks. The results clarify the critical relation between data, dimension, and neural network learning performance.
Conventional artificial neural networks are powerful tools in science and industry, but they can fail when applied to nonlinear systems where order and chaos coexist. We use neural networks that incorporate the structures and symmetries of Hamiltonian dynamics to predict phase space trajectories even as nonlinear systems transition from order to chaos. We demonstrate Hamiltonian neural networks on the canonical Henon-Heiles system, which models diverse dynamics from astrophysics to chemistry. The power of the technique and the ubiquity of chaos suggest widespread utility.
A finite element formulation is developed for a poroelastic medium consisting of an incompressible hyperelastic skeleton saturated by an incompressible fluid. The governing equations stem from mixture theory and the application is motivated by the study of interstitial fluid flow in brain tissue. The formulation is based on the adoption of an ALE perspective. We focus on a flow regime in which inertia forces are negligible. The stability and convergence of the formulation is discussed, and numerical results demonstrate agreement with the theory.
In the spirit of historic astronomical endeavors, we invited school groups across the globe to collaborate in a solar distance measurement using the rare June 5/6th transit of Venus. In total, we recruited 19 school groups spread over 6 continents and 10 countries to participate in our Hetu'u Global Network. Applying the methods of French astronomer Joseph-Nicolas Delisle, we used individual second and third Venus-Sun contact times to calculate the distance to the Sun. Ten of the sites in our network had amiable weather; 8 of which measured second contact and 5 of which measured third contact leading to consistent solar distance measurements of 152+/-30 million km and 163+/-30 million km respectively. The distance to the Sun at the time of the transit was 152.25 million km; therefore, our measurements are also consistent within 1sigma of the known value. The goal of our international school group network was to inspire the next generation of scientists using the excitement and accessibility of a rare astronomical event. In the process, we connected hundreds of participating students representing a diverse, multi-cultural group with differing political, economic, and racial backgrounds.