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 -
  Bingqing Cheng
DTSTART:20210112T131500Z
DTEND:20210112T141500Z
UID:TALK155041@talks.cam.ac.uk
CONTACT:Mateja Jamnik
DESCRIPTION:"Join us on Zoom":https://zoom.us/j/99166955895?pwd=SzI0M3pMVE
 kvNmw3Q0dqNDVRalZvdz09\n\nA central goal of computational physics and chem
 istry is to predict material properties using\nfirst principles methods ba
 sed on the fundamental laws of quantum mechanics. However\,\nthe high comp
 utational costs of these methods typically prevent rigorous predictions of
 \nmacroscopic quantities at finite temperatures\, such as chemical potenti
 al\, heat capacity and thermal conductivity.\n\nIn this talk\, I will firs
 t discuss how to enable such predictions by combining advanced statistical
  mechanics \nwith 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 t
 o excellent agreement with experiments\, but also reveals the\ncrucial rol
 e of nuclear quantum fluctuations in modulating the thermodynamic stabilit
 ies of\ndifferent phases of water. As another example [2]\, we simulated t
 he high pressure hydrogen system\nwith converged system size and simulatio
 n length\, and found\, contrary to established beliefs\,\nsupercritical be
 haviour of liquid hydrogen above the melting line. \nBesides the computati
 on of thermodynamic properties\, I will talk about transport properties:\n
 Ref [3] proposed a method to compute the heat conductivities of liquid jus
 t from equilibrium molecular dynamics trajectories.\n\nDuring the second p
 art of the talk\, I will rationalize why machine learning potentials work 
 at all\, and in particular\, the locality argument.\nI'll show that a mach
 ine-learning potential trained on liquid water alone can predict the prope
 rties of diverse ice phases\, \nbecause all the local environments charact
 erising the ice phases are found in liquid water [4].\n\nReferences:\n\n[1
 ] Bingqing Cheng\, Edgar A Engel\, Jörg Behler\, Christoph Dellago\, Mich
 ele Ceriotti. (2019) ab initio thermodynamics of liquid and solid water. P
 roceedings of the National Academy of Sciences\, 116 (4)\, 1110-1115.\n\n[
 2] Bingqing Cheng\, Guglielmo Mazzola\, Chris J. Pickard\, Michele Ceriott
 i. (2020) Evidence for supercritical behaviour of high-pressure liquid hyd
 rogen. Nature\, 585\, 217–220\n\n[3] Bingqing Cheng\, Daan Frenkel. (202
 0) Computing the Heat Conductivity of Fluids from Density Fluctuations. Ph
 ysical Review Letters\, 125\, 130602\n\n[4] Bartomeu Monserrat\, Jan Gerit
  Brandenburg\, Edgar A. Engel\, Bingqing Cheng. (2020) Liquid water contai
 ns the building blocks of diverse ice phases. Nature Communications 11.1: 
 1-8.
LOCATION:Zoom
END:VEVENT
END:VCALENDAR