BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Using Invariants and Metric Data Structures for Better Materials R
 esearch - Alya Bin Thabet - Montserrat Lab\, Cambridge
DTSTART:20221117T130000Z
DTEND:20221117T143000Z
UID:TALK179081@talks.cam.ac.uk
CONTACT:James Fergusson
DESCRIPTION:Designing representations of materials that are both accurate 
 and computationally tractable at scale is a challenging problem. Typical r
 epresentations\, such as graphs or arrays of structural and compositional 
 descriptors\, are inadequate because they are either incomplete or discont
 inuous under small perturbations.\nIn this talk\, I will introduce periodi
 c point sets with equivalence under isometry as models for finite and infi
 nite atomic structures. With this representation choice\, structural and c
 hemical isometry invariants provide metric measures of similarity between 
 all atomic structures. The invariants can also compensate for incomplete r
 epresentations and aid in studying perturbations.\nAfter demonstrating the
  invariants and their relevant properties\, I will describe the metric dat
 a structures and indexing method we now use and how they have improved our
  research workflow
LOCATION:Kavli Large Meeting Room\, Kavli Building
END:VEVENT
END:VCALENDAR