BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Drupal iCal API//EN
X-WR-CALNAME:Events items teaser
X-WR-TIMEZONE:America/Toronto
BEGIN:VTIMEZONE
TZID:America/Toronto
X-LIC-LOCATION:America/Toronto
BEGIN:DAYLIGHT
TZNAME:EDT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
DTSTART:20240310T070000
END:DAYLIGHT
BEGIN:STANDARD
TZNAME:EST
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
DTSTART:20231105T060000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
UID:69cf7d143ba37
DTSTART;TZID=America/Toronto:20240802T123000
SEQUENCE:0
TRANSP:TRANSPARENT
DTEND;TZID=America/Toronto:20240802T133000
URL:https://uwaterloo.ca/combinatorics-and-optimization/events/continuous-o
 ptimization-andersen-ang
SUMMARY:Continuous Optimization - Andersen Ang
CLASS:PUBLIC
DESCRIPTION:TITLE: Nonnegative Matrix Factorization in non-standard settin
 gs\, for\nfun\n\nSPEAKER:\n Andersen Ang\n\nAFFILIATION:\n University of S
 outhampton\n\nLOCATION:\n MC 6029\n\nABSTRACT: This abstract is broken in
 to 6 points.\n\n1. What is NMF: NMF is to find two elementwise nonnegative
  low-rank\nmatrices W and H such that M ≈ WH for a given elementwise\nno
 nnegative matrix M.\n\n2. NMF is commonly done in the Euclidean distance. 
 \n\n3. I argue that Euclidean distance is \"not good\"\, and a ray-to-ray\
 ndistance is better. \n\n4. Under L2-normalization onto the unit sphere\, 
 we arrive at the\ncosine angle distance\, which motivates the use of manif
 old\noptimization techniques. \n\n5. Under L1-normalization onto the simpl
 ex\, we arrive at the so-called\nAitchison geometry\, which contains funny
  algebra.\n\n6. Why do these: for curiosity and fun. For application\, the
 se\nnon-standard NMF can \"remove cloud\" from satellite images.
DTSTAMP:20260403T084052Z
END:VEVENT
END:VCALENDAR