BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//wp-events-plugin.com//7.2.3.1//EN
TZID:America/Los_Angeles
X-WR-TIMEZONE:America/Los_Angeles
BEGIN:VEVENT
UID:328@escience.washington.edu
DTSTART;TZID=America/Los_Angeles:20260312T163000
DTEND;TZID=America/Los_Angeles:20260312T172000
DTSTAMP:20260305T182736Z
URL:https://escience.washington.edu/events/uw-data-science-seminar-zepu-wa
 ng/
SUMMARY:UW Data Science Seminar: Zepu Wang
DESCRIPTION:Please join us for a UW Data Science Seminar featuring UW iUTS 
 lab PhD student Zepu Wang on Thursday\, March 12th from 4:30 to 5:20 p.m. 
 PT. The seminar will be held in IEB G109.\n"Decorrelating the Future: Join
 t Frequency Domain Learning for Spatio-temporal Forecasting"\nAbstract: St
 andard Direct Forecast models rely on point-wise objectives (e.g.\, Mean S
 quared Error) that overlook the complex spatio-temporal dependencies inher
 ent in graph signals. While recent frequency-domain methods like FreDF mit
 igate temporal autocorrelation\, they neglect spatial and cross-spatio-tem
 poral structures. To address this\, we propose FreST Loss (Frequency-enhan
 ced Spatio-Temporal Loss)\, which extends supervision to the joint spatio-
 temporal spectrum. By leveraging the Joint Spatio-temporal Fourier Transfo
 rm\, FreST Loss aligns predictions with ground truth in a unified spectral
  domain\, effectively decorrelating complex dependencies. Theoretical anal
 ysis confirms that this approach reduces the estimation bias in time-domai
 n training objectives. Extensive experiments on six real-world datasets de
 monstrate that FreST Loss is model-agnostic\, consistently improving state
 -of-the-art baselines by capturing holistic spatio-temporal dynamics.\n\nS
 peaker Bio: Zepu Wang is a second-year PhD student at the iUTS lab\, l
 ed by Prof. Xuegang (Jeff) Ban. He is an interdisciplinary AI researcher
  focusing on deep learning in time series and spatio-temporal data analysi
 s\, uncertainty quantification\, and their applications in transportation 
 and urban systems.\n\n\nThe 2025-2026 seminars will be held in person\, an
 d are free and open to the public.\n\n
ATTACH;FMTTYPE=image/jpeg:https://escience.washington.edu/wp-content/uploa
 ds/2026/03/Screen-Shot-2023-08-07-at-11.30.43-PM-e1772657949872.png
END:VEVENT
BEGIN:VTIMEZONE
TZID:America/Los_Angeles
X-LIC-LOCATION:America/Los_Angeles
BEGIN:DAYLIGHT
DTSTART:20260308T030000
TZOFFSETFROM:-0800
TZOFFSETTO:-0700
TZNAME:PDT
END:DAYLIGHT
END:VTIMEZONE
END:VCALENDAR