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:20200308T070000
END:DAYLIGHT
BEGIN:STANDARD
TZNAME:EST
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
DTSTART:20191103T060000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
UID:69f3f174b5f35
DTSTART;TZID=America/Toronto:20200813T160000
SEQUENCE:0
TRANSP:TRANSPARENT
DTEND;TZID=America/Toronto:20200813T160000
URL:https://uwaterloo.ca/statistics-and-actuarial-science/events/department
 -seminar-aaditya-ramdas-carnegie-mellon-university
SUMMARY:Department Seminar by Aaditya Ramdas\, Carnegie Mellon University
CLASS:PUBLIC
DESCRIPTION:CONCENTRATION INEQUALITIES FOR SAMPLING WITHOUT REPLACEMENT\, W
 ITH\nAPPLICATIONS TO POST-ELECTION AUDITS\n\n-------------------------\n\n
 Many practical tasks involve sampling sequentially without replacement\nfr
 om a finite population in order to estimate some parameter\, like a\nmean.
  We discuss how to derive powerful (new) concentration\ninequalities for t
 his setting using martingale techniques\, and apply\nit to auditing electi
 ons (see below).\n\nThis is joint work with my PhD student\, Ian Waudby-Sm
 ith\, who was an\nundergrad at UWaterloo. An early preprint is available h
 ere\n[https://arxiv.org/abs/2006.04347].\n\nMore details: When determining
  the outcome of an election\, electronic\nvoting machines are often employ
 ed for their tabulation speed and\ncost-effectiveness. Unlike paper ballot
 s\, these machines are\nvulnerable to software bugs and fraudulent tamperi
 ng. Post-election\naudits provide assurance that announced electoral outco
 mes are\nconsistent with paper ballots or voter-verifiable records. We pro
 pose\nan approach to election auditing based on confidence sequences\n(VAC
 SINE)—these are visualizable sequences of confidence sets for\nthe total
  number of votes cast for each candidate that adaptively\nshrink to zero w
 idth. These confidence sequences have uniform coverage\nfrom the beginning
  of an audit to the point of an exhaustive recount\,\nbut their main advan
 tage is that their error guarantee is immune to\ncontinuous monitoring and
  early stopping\, providing valid inference at\nany auditor-chosen\, data-
 dependent stopping time. We develop VACSINEs\nfor various types of electio
 ns including plurality\, approval\,\nranked-choice\, and score voting prot
 ocols.\n\nPLEASE NOTE: This talk will be given through Zoom. To join\, ple
 ase\nfollow this link: Department Seminar by Aaditya Ramdas\n[https://zoo
 m.us/j/8442836948?pwd=MVdCUFFCbVFuSzduQjhDQnNNZ3J1QT09#success].
DTSTAMP:20260501T001900Z
END:VEVENT
END:VCALENDAR