BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Machine Learning Like a Physicist
DTSTART:20181108T160000
DTEND:20181108T170000
DTSTAMP:20260429T022722Z
UID:1b6c2b9a4660c7bb1ec6d6e808947aafbe5c018f7349dc16fcced897
CATEGORIES:Conferences - Seminars
DESCRIPTION:Prof. Michele Ceriotti\nLaboratory for Computational Science a
 nd Modeling\, EPFL Lausanne\nChE-605 - Highlights in Energy Research semin
 ar series\nStatistical regression techniques have become very fashionable 
 as a tool to predict the properties of systems at the atomic scale\, sides
 tepping much of the computational cost of first-principles simulations and
  making it possible to perform simulations that require thorough statistic
 al sampling without compromising on the accuracy of the electronic structu
 re model. \nIn this talk I will argue how data-driven modelling can be ro
 oted in a mathematically rigorous and physically-motivated framework\, and
  how this is beneficial to the accuracy and the transferability of the mod
 el. I will also highlight how machine learning - despite amounting essenti
 ally at data interpolation - can provide important physical insights on th
 e behavior of complex systems\, on the synthesizability and on the structu
 re-property relations of materials.\nI will give examples concerning all s
 orts of atomistic systems\, from semiconductors to molecular crystals [1]\
 , and properties as diverse as drug-protein interactions [2]\, dielectric 
 response of aqueous systems[3] and NMR chemical shielding in the solid sta
 te [4].\n \n[1] F. Musil\, S. De\, J. Yang\, J. E. J. E. Campbell\, G. M.
  G. M. Day\, and M. Ceriotti\, Chem. Sci. 9 (2018) 1289\n[2] A. P. A. P. B
 artók\, S. De\, C. Poelking\, N. Bernstein\, J. R. J. R. Kermode\, G. Cs
 ányi\, and M. Ceriotti\, Sci. Adv. 3\, (2017) e1701816\n[3] A. Grisafi\, 
 D. M. Wilkins\, G. Csányi\, and M. Ceriotti\, Phys. Rev. Lett. 120 (2018)
  36002\n[4] http://shiftml.\n 
LOCATION:I17 4 K2 https://plan.epfl.ch/?room==I17%204%20K2
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR