BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Molecular Dynamics With On-The-Fly Machine Learning of QM forces
DTSTART:20150413T131500
DTEND:20150413T141500
DTSTAMP:20260411T071847Z
UID:e9e9256ac463390c13676edf9193acc98688437f5c7d933a164e58d7
CATEGORIES:Conferences - Seminars
DESCRIPTION:Prof. Alessandro De Vita\, Physics Department King's College L
 ondon UK\nAbstract: Many technologically important\, chemically complex ph
 enomena are currently beyond the reach of standard or O(N) first principle
 s molecular dynamics (MD) techniques\, mostly because the necessary model 
 system sizes are too large\, and/or the required simulation times are too 
 long. In most situations\, using classical MD is not a viable alternative\
 , as suitably general and accurate “reactive” force fields are not ava
 ilable\, nor are fitting databases a priori guaranteed to contain the info
 rmation necessary to describe all the chemical processes encountered along
  the dynamics. “QM/MM” techniques combining quantum and classical zone
 s in a single calculation are also not devoid of difficulties\, especially
  for applications on processes involving sustained mass transport into and
  out of the (e.g.\, fast moving) QM zone.\nIn this colloquium I will argue
  that this situation forces the use of MD techniques capable of incorporat
 ing accurate QM information generated at run time during the simulations [
 1\, 2]. It also effectively creates a novel market for dynamical databases
  coupled with specially-tuned Machine Learning (ML) force fields which min
 imise the computational workload by allowing QM subroutine calls only when
  “chemically novel” configurations are encountered along the system’
 s trajectory. I will present one such “Learn On the Fly” scheme\, effe
 ctively unifying First-Principles Molecular Dynamics and Machine Learning 
 into a single\, information efficient simulation scheme capable of learnin
 g/predicting atomic forces through Bayesian inference [3]. Interestingly\,
  QM-zone partitioning and execution via any of the existing O(N3) QM engin
 es is predicted to be a better option than using O(N) QM methods when deal
 ing with large QM zones in QM/MM calculations running on high-end parallel
  platforms [4-5].\nReferences\n[1] J.R.Kermode\, L.Ben-Bashat\, F.Atrash\,
  J.J.Cilliers\, D.Sherman and A.De Vita\, Nat. Commun. 4\, 2441 (2013).\n[
 2] A. Gleizer\, G. Peralta\, J. R. Kermode\, A. De Vita and D. Sherman\, P
 hys. Rev. Lett.\, 112\, 115501 (2014).\n[3] Z. Li\, J. R. Kermode and A. D
 e Vita\, Phys. Rev. Lett.\, 114\, 096405 (2015).\n[4] Cf.\, e.g.\, the US-
 DOE INCITE on SiO2 ML-Fracture Project https://www.alcf.anl.gov/projects/s
 io2-fracture-chemomechanics-machine-learning-hybrid-qmmm-scheme\n[5] M. Ca
 ccin\, Z. Li\, J. R. Kermode and A. De Vita\, in preparation.
LOCATION:MXF 1 https://plan.epfl.ch/?room==MXF%201
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR