BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:External FLAIR Seminar: Prof. Giulio Biroli
DTSTART:20220330T100000
DTSTAMP:20261006T035627Z
UID:2e7ca039c95b67262457b4b1b7dc7e07ed4fc683b7b66b0b9361c67f
CATEGORIES:Conferences - Seminars
DESCRIPTION:Giulio Biroli\nThe goal of external FLAIR seminars (which are 
 launched with this talk) is to have speakers from outside of EPFL presenti
 ng their work to all EPFL researchers interested by the theory of machine 
 learning. Please find below the information for the talk\, and do not hesi
 tate to share this announcement!\n\n \n\n \n\nTitle: Renormalization Gro
 up Theory and Machine Learning \n\n \n\nSpeaker: Giulio Biroli (ENS Pari
 s) \n\n\n \n\nAbstract: Reconstructing\, or generating\, high dimensiona
 l distributions starting from data is a central problem in machine learnin
 g and data sciences. \n\nI will present a method — The Wavelet Conditio
 nal Renormalization Group — that combines ideas from physics (renormaliz
 ation group theory) and computer science (wavelets\, stable representation
 s of operators). The Wavelet Conditional Renormalization Group allows to r
 econstruct in a very efficient way classes of high dimensional distributio
 ns hierarchically from large to small spatial scales. I will present the m
 ethod and then show its applications to data from statistical physics and 
 cosmology. The Wavelet Conditional Renormalization Group Method also provi
 des interesting insights on the interplay between structures of data and a
 rchitectures of deep neural networks. \n
LOCATION:GA 3 21 https://plan.epfl.ch/?room==GA%203%2021
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR