BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//https://techplay.jp//JP
CALSCALE:GREGORIAN
METHOD:PUBLISH
X-WR-CALDESC:[AI Security and Privacy Team Seminar] Talk by Prof. Kobbi Nis
 sim
X-WR-CALNAME:[AI Security and Privacy Team Seminar] Talk by Prof. Kobbi Nis
 sim
X-WR-TIMEZONE:Asia/Tokyo
BEGIN:VTIMEZONE
TZID:Asia/Tokyo
BEGIN:STANDARD
DTSTART:19700101T000000
TZOFFSETFROM:+0900
TZOFFSETTO:+0900
TZNAME:JST
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
UID:996246@techplay.jp
SUMMARY:[AI Security and Privacy Team Seminar] Talk by Prof. Kobbi Nissim
DTSTART;TZID=Asia/Tokyo:20260604T140000
DTEND;TZID=Asia/Tokyo:20260604T160000
DTSTAMP:20260526T101448Z
CREATED:20260526T060147Z
DESCRIPTION:イベント詳細はこちら\nhttps://techplay.jp/event/99624
 6?utm_medium=referral&utm_source=ics&utm_campaign=ics\n\nTitle: Protectin
 g the Undeleted in Machine Unlearning\nSpeaker：Prof. Kobbi Nissim (Geor
 getown University)\nDate and time: June 4 at 2:00 PM.\nOnline Venue (Zoom
 ): The URL will be provided only to registered participants.\n\nAbstract:
  Legal data protection standards such as the EU General Data Protection R
 egulation and the California Consumer Privacy Act give individuals the ri
 ght to request that their specific information be deleted\, also known as
  the Right to be Forgotten. This provision gave rise to machine unlearnin
 g\, a branch of machine learning focused on removing elements from traini
 ng data by efficiently producing a model that would have been obtained ha
 d the deleted data never been included\, namely\, “perfect retraining.
 ”\nIn this talk\, Prof. Nissim will discuss how data deletion affects p
 rivacy. He will first present a task that can be computed with strong pri
 vacy guarantees\, yet any perfect retraining mechanism for the task allow
 s an adversary controlling only a small number of data points to reconstr
 uct almost the entire dataset simply by issuing deletion requests.\nHe wi
 ll then discuss ways forward\, in particular a new cryptographically moti
 vated security definition that safeguards undeleted data points against l
 eakage caused by the deletion of other points. The talk will also show th
 at this definition permits several essential functionalities\, including 
 bulletin boards\, summations\, and statistical learning.\nThis is based o
 n joint work with Aloni Cohen\, Refael Kohen\, and Uri Stemmer.\n\nBio: P
 rof. Kobbi Nissim is McDevit Chair in Computer Science at Georgetown Univ
 ersity and is affiliated with Georgetown Law. His work focuses on the mat
 hematical formulation and understanding of privacy. His work with Dinur a
 nd Dwork in 2003 and 2004 initiated rigorous foundational research on pri
 vacy\, and in 2006 he introduced differential privacy with Dwork\, McSher
 ry\, and Smith. Prof. Nissim is a Fellow of the IACR. He received the Par
 is Kanellakis Theory and Practice Award in 2020\, the Caspar Bowden Award
  for Outstanding Research in Privacy Enhancing Technologies in 2019\, the
  Gödel Prize in 2017\, and test-of-time awards in 2013 and 2018.
LOCATION:オンライン
URL:https://techplay.jp/event/996246?utm_medium=referral&utm_source=ics&utm
 _campaign=ics
END:VEVENT
END:VCALENDAR