BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Robust Deep Learning Under Distribution Shift - Zack Lipton\, CMU
DTSTART:20200124T110000Z
DTEND:20200124T120000Z
UID:TALK137812@talks.cam.ac.uk
CONTACT:Adrian Weller
DESCRIPTION:We might hope that when faced with unexpected inputs\, well-de
 signed software systems would fire off warnings. However\, ML systems\, wh
 ich depend strongly on properties of their inputs (e.g. the i.i.d. assumpt
 ion)\, tend to fail silently. Faced with distribution shift\, we wish (i) 
 to detect and (ii) to quantify the shift\, and (iii) to correct our classi
 fiers on the fly—when possible. This talk will describe a line of recent
  work on tackling distribution shift. First\, I will focus on recent work 
 on label shift\, a more classic problem\, where strong assumptions enable 
 principled methods. Then I will discuss how recent tools from generative a
 dversarial networks have been appropriated (and misappropriated) to tackle
  dataset shift—characterizing and (partially) repairing a foundational f
 law in the method. Finally\, I will discuss new work that leverages human-
 in-the-loop feedback to develop classifiers that take into account causal 
 structure in text classification problem and appear (empirically) to benef
 it on a battery of out-of-domain evaluations.
LOCATION:Engineering Department\, CBL Room BE-438.
END:VEVENT
END:VCALENDAR