Machine learning the properties of water

Prof. Dr. Alexandre Reily Rocha, Instituto de Física Teórica, UNESP - São Paulo.

Water and its most common solid, ice I_h are essential for life on Earth, as well as for a variety of other Chemical and Physical processes. However, its microscopic structure can be deceivingly simple, and a fully atomistic understanding of its properties is still lacking. This is mostly because there is an interplay between long- and short-range interactions that is hard to capture using even the most up to date ab initio methodologies. In this talk I will discuss how we have been using machine learning methodologies, most notably neural networks to understand static and dynamic properties of water and ice Ih. I will discuss how to sample data to setup the best possible network. Finally, I will discuss the role of Nuclear Quantum Effects in both water and ice.

Instagram: https://instagram.com/ppfisufu?igshid...

Pagina do Programa de Pós-Graduação em Física
www.ppfis.infis.ufu.br