BEGIN:VCALENDAR
VERSION:2.0
CALSCALE:GREGORIAN
PRODID:iCalendar-Ruby
BEGIN:VEVENT
CATEGORIES:Conferences/Seminars/Lectures
DESCRIPTION:Speaker: Tess Smidt (MIT)\n\nTitle: Euclidean Symmetry Equi
variant Machine Learning for Atomic Systems --\n\n Overview\, Ap
plicatioins\, and Open Questions\n\n \n\nAbstract:\n\nAtomic systems (molec
ules\, crystals\, proteins\, etc.) are naturally represented by a set of co
ordinates in 3D space labeled by atom type. This is a challenging represent
ation to use for machine learning because the coordinates are sensitive to
3D rotations\, translations\, and inversions (the symmetries of 3D Euclidea
n space). In this talk I’ll give an overview of Euclidean invariance and eq
uivariance in machine learning for atomic systems. Then\, I’ll share some r
ecent applications of these methods on a variety of atomistic modeling task
s (ab initio molecular dynamics\, prediction of crystal properties\, and sc
aling of electron density predictions). Finally\, I’ll explore open questio
ns in expressivity\, data-efficiency\, and trainability of methods leveragi
ng invariance and equivariance.
DTEND:20230510T211500Z
DTSTAMP:20241105T044028Z
DTSTART:20230510T201500Z
LOCATION:MIT-Math Dept.\, Room 2-449
SEQUENCE:0
SUMMARY:Seminar: Numerical Methods for Partial Differential Equations
UID:tag:localist.com\,2008:EventInstance_43155175404936
URL:https://calendar.mit.edu/event/numerical_methods_for_partial_differenti
al_equations_5638
END:VEVENT
END:VCALENDAR