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:20180311T070000
END:DAYLIGHT
BEGIN:STANDARD
TZNAME:EST
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
DTSTART:20171105T060000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
UID:69b4df99b7aab
DTSTART;TZID=America/Toronto:20180925T130000
SEQUENCE:0
TRANSP:TRANSPARENT
DTEND;TZID=America/Toronto:20180925T130000
URL:https://uwaterloo.ca/artificial-intelligence-group/events/phd-defence-l
 earning-sparse-orthogonal-wavelet-filters
LOCATION:DC - William G. Davis Computer Research Centre 200 University Aven
 ue West 2310 Waterloo ON N2L 3G1 Canada
SUMMARY:PhD Defence: Learning Sparse Orthogonal Wavelet Filters
CLASS:PUBLIC
DESCRIPTION:Daniel Recoskie\, PhD candidate\nDavid R. Cheriton School of Co
 mputer Science\n\nThe wavelet transform is a well-studied and understood a
 nalysis\ntechnique used in signal processing. In wavelet analysis\, signal
 s are\nrepresented by a sum of self-similar wavelet and scaling functions.
 \nTypically\, the wavelet transform makes use of a fixed set of wavelet\nf
 unctions that are analytically derived. We propose a method for\nlearning 
 wavelet functions directly from data. We impose an\northogonality constrai
 nt on the functions so that the learned wavelets\ncan be used to perform b
 oth analysis and synthesis. We accomplish this\nby using gradient descent 
 and leveraging existing automatic\ndifferentiation frameworks. Our learned
  wavelets are able to capture\nthe structure of the data by exploiting spa
 rsity. We show that the\nlearned wavelets have similar structure to tradit
 ional wavelets.
DTSTAMP:20260314T041001Z
END:VEVENT
END:VCALENDAR