BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Bohnenblust--Hille inequalities and low-degree learning - Haonan Z
 hang (University of South Carolina)
DTSTART:20241206T111000Z
DTEND:20241206T115000Z
UID:TALK218662@talks.cam.ac.uk
DESCRIPTION:A fundamental problem from computational learning theory is to
  efficiently reconstruct an unknown Boolean function. One classical result
  of this problem for the random query model is the low-degree algorithm of
  Linial\, Mansour\, and Nisan in 1993. This method saw exponential improve
 ment in 2022 by Eskenazis and Ivanisvili via a family of dimension-free po
 lynomial inequalities named after Bohnenblust and Hille dating back to Lit
 tlewood's work in 1930. In this talk\, I will review the recent progress a
 long this line of research on discrete quantum systems\, and discuss paral
 lel advances in classical harmonic analysis. This is based on joint work w
 ith Alexander Volberg and Joseph Slote.
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR