BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:What is an information projection? An invitation to information ge
 ometry
DTSTART:20170427T140000
DTEND:20170427T160000
DTSTAMP:20261005T001431Z
UID:2b8cf0408584dd6cacc624e702b628d1e7f45bed7cd91a61a8e80e6a
CATEGORIES:Conferences - Seminars
DESCRIPTION:Frank Nielsen\, Professeur at Ecole Polytechnique\, France.\nh
 ttps://www.lix.polytechnique.fr/~nielsen/\n\nFrank Nielsen received his Ph
 D (1996) and his habilitation (2006) on computational geometry\nfrom the u
 niversity of Nice-Sophia Antipolis\, France.\nAfter the french national se
 rvice\, he joined  Sony CSL (Japan) in 1997.\nHe is currently  professor
  in the computer science department of Ecole polytechnique (France).\nHe c
 o-organizes with Frédéric Barbaresco (Thales) the biannual Geometric Sci
 ences of Information (GSI\, gsi2017.org) conference\,\nand is currently co
 -editor of the newly launched Springer journal of Information Geometry and
 \nan associate editor of MDPI Entropy.\nIn this tutorial\, we first presen
 t two renown statistical inference methods:\nThe Maximum Likelihood Estima
 tor and the Maximum Entropy Principle.\nWe show how these estimators can b
 e modeled as Kullback-Leibler divergence minimization problems\,\nand ther
 efore interpreted geometrically as information projections.\nWe then prese
 nt the basic construction of dually flat Shannon spaces and describe some 
 useful concepts and tools\n (eg.\, space of spheres\, statistical Voronoi
  diagrams).\nFinally\, we  touch upon the role of divergences in informat
 ion theory\, statistics\, pattern recognition and machine learning.\n 
LOCATION:INR113
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR