BEGIN:VCALENDAR
CALSCALE:GREGORIAN
PRODID:iCalendar-Ruby
VERSION:2.0
BEGIN:VEVENT
DESCRIPTION: We study the problem of computing a point in the convex hull\,
  and the related problem of computing a separating hyperplane\,  under the 
 constraint of differential privacy . Intuitively\, differential privacy mea
 ns that our output should be robust to small changes in the input (for exam
 ple to adding or deleting a point). We study the minimum size of the input 
 needed to achieve such a private computation (sample complexity) and its ti
 me complexity.  Even in one dimension the problem is non-trivial and we wil
 l first focus on this case. Several interesting open problems will be prese
 nted as well.    No previous background on differential privacy will be ass
 umed. 
DTSTAMP:20200121T160800
DTSTART:20200127T141500
CLASS:PUBLIC
LOCATION:Freie Universität Berlin \n Institut für Informatik \n Takustr. 9 
 \n 14195 Berlin \n Room 005 (Ground Floor)
SEQUENCE:0
SUMMARY:Haim Kaplan (Tel Aviv University): Privately Learning Thresholds
UID:95692972@www.facetsofcomplexity.de
URL:http://www.facetsofcomplexity.de/monday/20200127-L-Kaplan.html
END:VEVENT
END:VCALENDAR