BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Predicting material properties with the help of machine learning -
  Dr Bingqing Cheng\, University of Cambridge
DTSTART:20210127T143000Z
DTEND:20210127T153000Z
UID:TALK154219@talks.cam.ac.uk
CONTACT:Lisa Masters
DESCRIPTION:A central goal of computational physics and chemistry is to pr
 edict material properties using\nfirst principles methods based on the fun
 damental laws of quantum mechanics. However\,\nthe high computational cost
 s of these methods typically prevent rigorous predictions of\nmacroscopic 
 quantities at finite temperatures\, such as chemical potential\, heat capa
 city and thermal conductivity.\n\nIn this talk\, I will first discuss how 
 to enable such predictions by combining advanced statistical mechanics \nw
 ith data-driven machine learning interatomic potentials. As an example [1]
 \, for the\nomnipresent and technologically essential system of water\, a 
 first-principles thermodynamic\ndescription not only leads to excellent ag
 reement with experiments\, but also reveals the\ncrucial role of nuclear q
 uantum fluctuations in modulating the thermodynamic stabilities of\ndiffer
 ent phases of water. As another example [2]\, we simulated the high pressu
 re hydrogen system\nwith converged system size and simulation length\, and
  found\, contrary to established beliefs\,\nsupercritical behaviour of liq
 uid hydrogen above the melting line. \nBesides the computation of thermody
 namic properties\, I will talk about transport properties:\nRef [3] propos
 ed a method to compute the heat conductivities of liquid just from equilib
 rium molecular dynamics trajectories.\n\nDuring the second part of the tal
 k\, I will rationalize why machine learning potentials work at all\, and i
 n particular\, the locality argument.\nI'll show that a machine-learning p
 otential trained on liquid water alone can predict the properties of diver
 se ice phases\, \nbecause all the local environments characterising the ic
 e phases are found in liquid water [4].\n\nReferences:\n[1] Bingqing Cheng
 \, Edgar A Engel\, Jörg Behler\, Christoph Dellago\, Michele Ceriotti. (2
 019) ab initio thermodynamics of liquid and solid water. Proceedings of th
 e National Academy of Sciences\, 116 (4)\, 1110-1115.\n[2] Bingqing Cheng\
 , Guglielmo Mazzola\, Chris J. Pickard\, Michele Ceriotti. (2020) Evidence
  for supercritical behaviour of high-pressure liquid hydrogen. Nature\, 58
 5\, 217–220\n[3] Bingqing Cheng\, Daan Frenkel. (2020) Computing the Hea
 t Conductivity of Fluids from Density Fluctuations. Physical Review Letter
 s\, 125\, 130602\n[4] Bartomeu Monserrat\, Jan Gerit Brandenburg\, Edgar A
 . Engel\, Bingqing Cheng. (2020) Liquid water contains the building blocks
  of diverse ice phases. Nature Communications 11.1: 1-8.
LOCATION:Zoom: Meeting ID: 976 0491 4122 Passcode: 108408
END:VEVENT
END:VCALENDAR