BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:CIS - "Get to know your neighbors" Seminar series - Prof. Lenka Zd
 eborova
DTSTART:20201109T151500
DTEND:20201109T161500
DTSTAMP:20260428T051707Z
UID:b06a4e0b6d3eb62949e4fe2d68a737c3fedfd40ab88da3489e1d44ba
CATEGORIES:Conferences - Seminars
DESCRIPTION:Prof. Lenka Zdeborova\nTitle: Insights on gradient-based algor
 ithms in high-dimensional non-convex learning.\n\nAbstract: Gradient desce
 nt algorithms and their noisy variants\, such as the Langevin dynamics or 
 multi-pass SGD\, are the working horse of machine learning. Yet their beha
 viour and performance remain perplexing\, in particular in the high-dimens
 ional non-convex setting. In this talk\, I will highlight the importance o
 f the associated theoretical questions. I will then present several high-d
 imensional and non-convex statistical learning problems in which the perfo
 rmance of gradient-based algorithms can be analysed down to a constant. Th
 e common point of these settings is that the data come from a probabilisti
 c generative model leading to problems for which\, in the high-dimensional
  limit\, statistical physics provides exact closed solutions for the perfo
 rmance of the gradient-based algorithms. The covered settings include the 
 spiked mixed matrix-tensor model and the phase retrieval.\n\nBio: (subset 
 of) Lenka Zdeborová is a Professor of Physics and of Computer Science in 
 École Polytechnique Fédérale de Lausanne. She received a PhD in physics
  from University Paris-Sud and from Charles University in Prague in 2008. 
 She spent two years in the Los Alamos National Laboratory as the Director'
 s Postdoctoral Fellow. Between 2010 and 2020 she was a researcher at CNRS 
 working in the Institute of Theoretical Physics in CEA Saclay\, France. In
  2014\, she was awarded the CNRS bronze medal\, in 2016 Philippe Meyer pri
 ze in theoretical physics and an ERC Starting Grant\, in 2018 the Irène J
 oliot-Curie prize. She is an editorial board member for Journal of Physics
  A\, Physical Review E\, Physical Review X\, SIMODS\, Machine Learning: Sc
 ience and Technology\, and Information and Inference. Lenka's expertise is
  in applications of methods developed in statistical physics\, such as adv
 anced mean field methods\, replica method and related message-passing algo
 rithms\, to problems in machine learning\, signal processing\, inference a
 nd optimization. \n\nThe Center for Intelligent Systems at EPFL (CIS) i
 s a collaboration among IC\, SB\, and STI that brings together researchers
  working on different aspects of Intelligent Systems.\n \nIn order to pro
 mote exchanges among researchers and encourage the creation of new\, colla
 borative projects\, CIS is organizing a "Get to know your neighbors" serie
 s. Each seminar will consist of 1-2 short overview presentations geared to
  the general public at EPFL.\n \nMonday\, 9th November 2020 from 3:15 to 
 4:15 pm\n\nNB: Video recordings of the seminars will be made available on 
 our website and published on our social media pages
LOCATION:A définir https://go.epfl.ch/CIS_GTKYN_L_Zdeborova https://go.ep
 fl.ch/CIS_GTKYN_L_Zdeborova
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR