BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//https://techplay.jp//JP
CALSCALE:GREGORIAN
METHOD:PUBLISH
X-WR-CALDESC:差分プライバシーを超えたデータ保護 [AIセキ
 ュリティ&プライバシーチーム]
X-WR-CALNAME:差分プライバシーを超えたデータ保護 [AIセキ
 ュリティ&プライバシーチーム]
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:982160@techplay.jp
SUMMARY:差分プライバシーを超えたデータ保護 [AIセキュリ
 ティ&プライバシーチーム]
DTSTART;TZID=Asia/Tokyo:20250624T103000
DTEND;TZID=Asia/Tokyo:20250624T113000
DTSTAMP:20260528T014706Z
CREATED:20250609T140143Z
DESCRIPTION:イベント詳細はこちら\nhttps://techplay.jp/event/98216
 0?utm_medium=referral&utm_source=ics&utm_campaign=ics\n\n\n\nAbstract\n\n
 Running machine learning and AI algorithms on personal and sensitive data
  raises privacy concerns and creates the potential for inadvertent inform
 ation leakage. For example\, text messages or images can be extracted fro
 m generative models. However\, analyzing such data can significantly bene
 fit individuals and society\, particularly in areas like healthcare and t
 ransportation. To balance these conflicting objectives\, it's essential t
 o deploy and securely implement data analysis methods with strong confide
 ntiality guarantees.\n\nIn this talk\, I will explore the challenges and 
 opportunities in achieving this goal. I'll start by detailing attacks tha
 t target not only machine learning algorithms but also naive implementati
 ons of algorithms that offer rigorous theoretical guarantees\, such as di
 fferential privacy. Following this\, I will discuss approaches to mitigat
 e these attack vectors\, including property-preserving data analysis. Spe
 cifically\, I will present our work on protecting dataset properties that
  extend beyond traditional record-level privacy—for instance\, safeguar
 ding subpopulation information instead of focusing solely on individual r
 ecords. Finally\, I will introduce ElephantDP\, a system designed to prov
 ide strong security guarantees even when differential privacy algorithms 
 are executed in insecure environments.\n\n\n\nBio\n\nOlya Ohrimenko is a 
 Professor at The University of Melbourne\, where she joined in 2020 after
  six years at Microsoft Research in Cambridge\, UK. Her research focuses 
 on the privacy and integrity of machine learning and AI algorithms\, data
  analysis tools\, and cloud computing. She works on various topics includ
 ing:\n\n\nDifferential privacy\nDataset confidentiality\nVerifiable and d
 ata-oblivious computation\nTrusted execution environments\nSide-channel a
 ttacks and their mitigations\n\n\nOlya has collaborated with organization
 s such as the Australian Bureau of Statistics\, National Australia Bank\,
  and Microsoft. She has also secured solo and joint research grants from 
 Meta\, Oracle\, and the Australian Department of Defence. Her contributio
 ns have been recognized with a Commendation for Outstanding Research Cont
 ribution in the 2025 CORE Awards. She was also a finalist in the AI in Cy
 ber Security category of the Women in AI Asia-Pacific Awards in both 2023
  and 2024.\n\nFor more information\, please visit https://oohrimenko.gith
 ub.io.
LOCATION:オンライン
URL:https://techplay.jp/event/982160?utm_medium=referral&utm_source=ics&utm
 _campaign=ics
END:VEVENT
END:VCALENDAR