BEGIN:VCALENDAR
VERSION:2.0
PRODID:icalendar-ruby
CALSCALE:GREGORIAN
X-WR-CALNAME:NA&PDE seminar: Dr. Daniel Hayes
X-WR-TIMEZONE:Eastern Time (US & Canada)
BEGIN:VEVENT
DTSTAMP:20260513T224523Z
UID:tag:localist.com\,2008:EventInstance_52182462246398
DTSTART:20260225T151500Z
DTEND:20260225T161500Z
DESCRIPTION:Title: Tensor Train and Cross Approximations for Large Scale Da
 ta Compression\n\n\n\nAbstract: In many areas of scientific computing\, st
 atistical analysis\, and machine learning\, the ability to efficiently and
  accurately handle large scale and high dimensional data is a rapidly grow
 ing necessity. Many powerful techniques have been developed to approach th
 is task\, one being Tensor-Train (TT) and variants such as Tensor-Train Ad
 aptive Cross Approximation (TTACA). TTACA provides a framework suited for 
 working with extremely large datasets while maintaining low memory and com
 putational complexity. In this talk\, I will introduce the foundations at 
 the matrix level and the base TTACA algorithm. I will then present two pro
 jects I have worked on: (1) the development of a subtensor parallel implem
 entation of TTACA that utilizes more computing resources to increase effic
 iency\, and (2) a post-processing oversampling algorithm designed to enhan
 ce approximation accuracy. Numerical experiments demonstrate improved comp
 utational efficiency through subtensor parallelism and systematic error re
 duction via oversampling.
GEO:39.681267;-75.754997
LOCATION:Ewing Hall\, 336
SUMMARY:NA&PDE seminar: Dr. Daniel Hayes
URL;VALUE=URI:https://events.udel.edu/event/napde-seminar-dr-daniel-hayes-4
 306
CATEGORIES:Academics
CATEGORIES:College of Arts and Sciences
END:VEVENT
END:VCALENDAR