BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Robustness\, Stability and the pNML method
DTSTART:20190719T130000
DTEND:20190719T150000
DTSTAMP:20260407T131948Z
UID:81f25482e5132839d45a63ab0b8d5f24035e1df2e3792029531cff97
CATEGORIES:Conferences - Seminars
DESCRIPTION:Andreas Maggiori\nEDIC candidacy exam\nExam president: Prof. M
 ichael Kapralov\nThesis advisor: Prof. Ruediger Urbanke\nThesis co-advisor
 : Prof. Ola Svensson\nCo-examiner: Dr. Olivier Lévêque\n\n\nAbstract\nTr
 aditional approaches fail to explain the recent success story of machine l
 earning. Furthermore\, common practices in ML are lacking a deep theoretic
 al understanding. In this proposal\, we investigate the relation between t
 he generalization of a learning algorithm\, its robustness\, and its stabi
 lity. Moreover\, we present an alternative supervised learning framework w
 here the goal is to obtain worst case guarantees. The objective is to expl
 ore the connections between these notions and apply similar techniques to 
 further analyze and extend common ML algorithms.\n\nBackground papers\nRob
 ustness and Generalization\, by Huan Xu\, Shie Mannor.\nTrain faster\, gen
 eralize better: Stability of stochastic gradient descent\, by Moritz Hardt
 \, Benjamin Recht\, Yoram Singer.\nUniversal Supervised Learning for Indiv
 idual Data\, by Yaniv Fogel\, Meir Feder.\n\n 
LOCATION:INR 113 https://plan.epfl.ch/?room=INR113
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR