BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Asymptotics and Dynamics of Learning\,
DTSTART:20230616T130000
DTEND:20230616T150000
DTSTAMP:20260408T134458Z
UID:ccd4660a4b7d3b9554cc88abe3a94dc578e1418c14cd3fd2d720f191
CATEGORIES:Conferences - Seminars
DESCRIPTION:Yatin Dandi\nEDIC candidacy exam\nExam president: Prof. Nicola
 s Flammarion\nThesis advisor: Prof. Lenka Zdeborova\nCo-examiner: Prof. L
 énaïc Chizat\n\nAbstract\nTheoretical analysis of the asymptotics of sys
 tems involving a large number of interacting entities has enabled massive 
 progress in fields ranging from statistical physics\, high dimensional sta
 tistics\, combinatorial optimization\, to the theory of machine learning. 
 In this PhD\, we aim to build upon the techniques developed in the above f
 ields to obtain novel insights into the phase transitions and dynamics in 
 high-dimensional computational phenomenon\, particularly for domains such 
 as representation learning and causality\, where many central questions re
 main unanswered.\n\nBackground papers\nHigh-dimensional Asymptotics of Fea
 ture Learning: How One Gradient Step Improves the Representation: https:/
 /openreview.net/pdf?id=akddwRG6EGi\nOnline stochastic gradient descent on 
 non-convex losses from high-dimensional inference: https://jmlr.csail.mit
 .edu/papers/volume22/20-1288/20-1288.pdf\nMajority dynamics on trees and t
 he dynamic cavity method: https://arxiv.org/abs/0907.0449\, https://proje
 cteuclid.org/journals/annals-of-applied-probability/volume-21/issue-5/Majo
 rity-dynamics-on-trees-and-the-dynamic-cavity-method/10.1214/10-AAP729.ful
 l
LOCATION:BC 329 https://plan.epfl.ch/?room==BC%20329
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR