BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Department of Statistics - ECPv5.1.6//NONSGML v1.0//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
X-WR-CALNAME:Department of Statistics
X-ORIGINAL-URL:https://stat.wisc.edu
X-WR-CALDESC:Events for Department of Statistics
BEGIN:VTIMEZONE
TZID:America/Chicago
BEGIN:DAYLIGHT
TZOFFSETFROM:-0600
TZOFFSETTO:-0500
TZNAME:CDT
DTSTART:20200308T080000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0500
TZOFFSETTO:-0600
TZNAME:CST
DTSTART:20201101T070000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTART;TZID=America/Chicago:20200214T160000
DTEND;TZID=America/Chicago:20200214T170000
DTSTAMP:20200928T002128
CREATED:20200113T170915Z
LAST-MODIFIED:20200207T161514Z
UID:2733-1581696000-1581699600@stat.wisc.edu
SUMMARY:Statistics Seminar
DESCRIPTION:Title: Differentially Private Inference for Binomial Data \nPresenter: Jordan Awan PhD Student Pennsylvania State University \nAbstract: Differential Privacy (DP) is a rigorous framework which quantifies the disclosure risk of statistical procedures computed on sensitive data. DP methods/mechanisms require the introduction of additional randomness beyond sampling in order to limit the disclosure risk. However\, their implementations often introduce excessive noise\, reducing the utility and validity of statistical results\, especially in small samples. \n In this paper\, we first derive uniformly most powerful (UMP) tests for simple and one-sided hypotheses for a population proportion within the framework of DP\, optimizing finite sample performance. By studying the privacy constraints\, we prove a `Neyman-Pearson Lemma’ for binomial data under DP\, resulting in a DP-UMP test. Our tests can also be stated as a post-processing of a DP summary statistic\, whose distribution we coin “Truncated-Uniform-Laplace” (Tulap)\, a generalization of the Staircase and discrete Laplace distributions. \n We show that by post-processing the Tulap statistic\, we are able to obtain exact p-values corresponding to the DP-UMP\, uniformly most accurate (UMA) one-sided confidence intervals\, uniformly most powerful unbiased (UMPU) two-sided tests\, and uniformly most accurate unbiased (UMAU) two-sided confidence intervals. As each of these quantities are a post-processing of the same summary statistic\, there is no increased privacy risk by including these additional results\, allowing for a complete statistical analysis at a fixed privacy cost. \nJordan Awan Seminar
URL:https://stat.wisc.edu/event/statistics-seminar-4-2020-02-14/
LOCATION:133 SMI
END:VEVENT
END:VCALENDAR