BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Topological analysis in information spaces
DTSTART:20161124T101500
DTEND:20161124T113000
DTSTAMP:20260407T064249Z
UID:52d771cf43e75e34dfdcfed2f168bb92fb22d83af2a4be5dea880761
CATEGORIES:Conferences - Seminars
DESCRIPTION:Hubert Wagner (IST Vienna)\nUnderstanding high dimensional dat
 a remains a challenging problem.\nComputational topology\, in an effort du
 bbed Topological Data Analysis\n(TDA)\, promises to simplify\, characteriz
 e and compare such\ndata. However\, TDA focuses on Euclidean spaces\, whil
 e many types of\nhigh-dimensional data naturally live in non-Euclidean one
 s. Spaces\nderived from text\, speech\, image... data are best characteriz
 ed by\nnon-metric dissimilarities\, many of which are inspired by\ninforma
 tion-theoretical concepts.  Such spaces will be called\ninformation space
 s.\n\nI will present the theoretical foundations of topological analysis i
 n\ninformation spaces. First\, intuition behind basic computational\ntopol
 ogy methods is given. Then\, various dissimilarity measures are\ndefined a
 long with information theoretical and geometric\ninterpretation. Finally\,
  I will show that the framework of TDA can be\nextended to information spa
 ces and discuss the implications.\n\nNo previous knowledge about (computat
 ional) topology or information\ntheory is required. This is joint work wit
 h Herbert Edelsbrunner and\nZiga Virk. We look for interesting\, practical
  applications for the\nabove!
LOCATION:MA 31
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR