BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:A data-derived approach to initialise densities in DFT - Andrew Fo
 wler
DTSTART:20190130T120000Z
DTEND:20190130T123000Z
UID:TALK119473@talks.cam.ac.uk
CONTACT:Nick Woods
DESCRIPTION:Standard methods to initialise densities in DFT are based upon
  \nnon-interacting contributions from atoms that are assumed to be in \nva
 cuum. We apply standard techniques from total energy methods to \ninterpol
 ate ground state densities.\n\nBy using a non-Bayesian ensemble method to 
 quantify uncertainty\, we \napply only densities that are likely to be acc
 urate to initialise KS DFT \ncalculations. We show for a specific system\,
  that data-derived over \nstandard initial densities can accelerate conver
 gence to self \nconsistency by up to 67%.
LOCATION:TCM Seminar Room\, Cavendish Laboratory
END:VEVENT
END:VCALENDAR