BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Date iCal//NONSGML kigkonsult.se iCalcreator 2.20.2//
METHOD:PUBLISH
BEGIN:VTIMEZONE
TZID:America/New_York
BEGIN:STANDARD
DTSTART:20231105T020000
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20240310T020000
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
UID:calendar.2740.events_uoft_date.0@www.statistics.utoronto.ca
CREATED:20231030T135131Z
DESCRIPTION:\nWhen and Where: \nThursday, November 23, 2023 3:30 pm to 4:
 30 pm \n 9014 \n Ontario Power Building \n 700 University Avenue, Toronto
 , ON M5G 1Z5 \n\nSpeakers \nHan Zhao, University of Illinois at Urbana-C
 hampaign \n\nDescription: \nTo mitigate the bias exhibited by machine lear
 ning models, fairness criteria can be integrated into the training proces
 s to ensure fair treatment across all demographics, but it often comes at
  the expense of model performance. Understanding such tradeoffs, therefor
 e, underlies the design of optimal and fair algorithms. In this talk, I 
 will first discuss our recent work on characterizing the inherent tradeoff
  between fairness and accuracy in classification problems, where we show 
 that the cost of fairness could be characterized by the optimal value of a
  Wasserstein-barycenter problem. Then I will show that the complexity of l
 earning the optimal fair predictor is the same as learning the Bayes predi
 ctor, and present a post-processing algorithm based on the solution to th
 e Wasserstein-barycenter problem that derives the optimal fair predictors 
 from Bayes score functions. I will also present the empirical results of o
 ur fair algorithm and conclude the talk with some discussion on the close 
 interplay between algorithmic fairness and domain generalization.Please jo
 in the event.About Han ZhaoDr. Han Zhao is an Assistant Professor of Compu
 ter Science at the University of Illinois Urbana-Champaign (UIUC). He is a
 lso an Amazon Visiting Academic at Amazon AI and Search Science. Dr. Zhao 
 earned his Ph.D. degree from Carnegie Mellon University under the guidance
  of Prof. Geoff Gordon. His research interest is centered around trustwort
 hy machine learning, with a focus on transfer learning, domain adaptatio
 n/generalization, and algorithmic fairness. His long-term goal is to buil
 d trustworthy ML systems that are efficient, robust, fair, private, an
 d interpretable. He received his bachelor's degree in Computer Science fro
 m Tsinghua University and his master's degree in mathematics from the Univ
 ersity of Waterloo. \n700 University Avenue, Toronto, ON M5G 1Z5 \n\nCat
 egories \n Seminar Series \n\nAudiences \n FacultyGraduate Students
DTSTART;TZID=America/New_York:20231123T153000
DTEND;TZID=America/New_York:20231123T163000
LAST-MODIFIED:20231030T135428Z
LOCATION:700 University Avenue, Toronto, ON M5G 1Z5
SUMMARY:Fair and Optimal Prediction via Post-Processing
URL;TYPE=URI:https://www.statistics.utoronto.ca/events/fair-and-optimal-pre
 diction-post-processing
END:VEVENT
END:VCALENDAR