BEGIN:VCALENDAR
VERSION:2.0
PRODID:researchseminars.org
CALSCALE:GREGORIAN
X-WR-CALNAME:researchseminars.org
BEGIN:VEVENT
SUMMARY:Ila Varma
DTSTART:20260930T123000Z
DTEND:20260930T133000Z
DTSTAMP:20261026T105426Z
UID:IpmMathColloquium/1
DESCRIPTION:Title: <a href="https://researchseminars.org/talk/IpmMathCollo
 quium/1/">Counting number fields and predicting asymptotics</a>\nby Ila Va
 rma as part of IPM Mathematics Colloquium\n\nAbstract: TBA\n
LOCATION:https://researchseminars.org/talk/IpmMathColloquium/1/
END:VEVENT
BEGIN:VEVENT
SUMMARY:Shayan Oveis Gharan (Professor of Computer Science & Engineering)
DTSTART:20261028T140000Z
DTEND:20261028T150000Z
DTSTAMP:20261026T105426Z
UID:IpmMathColloquium/2
DESCRIPTION:Title: <a href="https://researchseminars.org/talk/IpmMathCollo
 quium/2/">High Dimensional Expanders and Separated Sets</a>\nby Shayan Ove
 is Gharan (Professor of Computer Science & Engineering) as part of IPM Mat
 hematics Colloquium\n\n\nAbstract\nAbstract: Given a set of unit vectors v
 1\,...\,vn (on a d-dimensional sphere). If they are pairwise orthogonal th
 en we can cluster them (into n clusters) such that the distance between an
 y two clusters is at least say 1. Now\, suppose we have a weaker assumptio
 n that the average absolute value of the inner product |<vi\,vj>| (for i\,
 j chosen uniformly at random) is 1/n^0.99. Can we still find (at least) 2 
 clusters (of linear size) with constant distance?\nThe problem is closely 
 related to analysis of random walks and approximation algorithms for the s
 parsest cut problem. I will show that contrary to our belief the answer is
  no. The proof is based on the rapidly evolving area theory of high dimens
 ional expanders.\nBased on a joint work with Farzam Ebrahimnejad.\n\nJoin 
 Zoom Meeting:\n\nhttps://us06web.zoom.us/j/86258815402?pwd=6vkzDDvvk49gp5u
 JO4358acnj4k8Yb.1\n\nor \n\nhttps://zoom.us/join\n\nMeeting ID: 862 5881 5
 402\nPasscode: 260930\n
LOCATION:https://researchseminars.org/talk/IpmMathColloquium/2/
END:VEVENT
END:VCALENDAR