BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//talks.cam.ac.uk//v3//EN
BEGIN:VTIMEZONE
TZID:Europe/London
BEGIN:DAYLIGHT
TZOFFSETFROM:+0000
TZOFFSETTO:+0100
TZNAME:BST
DTSTART:19700329T010000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0100
TZOFFSETTO:+0000
TZNAME:GMT
DTSTART:19701025T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
CATEGORIES:DAMTP Statistical Physics and Soft Matter Seminar
SUMMARY:Renormalisation group and machine learning: the Wa
 velet-Conditional RG  - Giulio Biroli (ENS Paris)
DTSTART;TZID=Europe/London:20230516T130000
DTEND;TZID=Europe/London:20230516T140000
UID:TALK199942AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/199942
DESCRIPTION:Reconstructing\, or generating\, high dimensional 
 distributions starting from data is a central prob
 lem in machine learning and data sciences.\nI will
  present a method —The Wavelet Conditional Renorma
 lization Group —that combines ideas from physics (
 renormalization group theory) and computer science
  (wavelets\, stable representations of operators).
  The Wavelet Conditional Renormalization Group all
 ows to reconstruct in a very efficient way classes
  of high dimensional probability distributions hie
 rarchically from large to small spatial scales\, a
 nd to perform RG directly from data.  It allows to
  bridge the gap between approaches based on physic
 al intuition and modern machine learning algorithm
 s. I will present the method and then show its app
 lications to data from statistical physics and cos
 mology. I shall also discuss the interesting insig
 hts that our method offers on the interplay betwee
 n structures of data and architectures of deep neu
 ral networks.\n
LOCATION:Center for Mathematical Sciences\, Lecture room MR
 4
CONTACT:Sarah Loos
END:VEVENT
END:VCALENDAR