BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//talks.cam.ac.uk//v3//EN
BEGIN:VTIMEZONE
TZID:Europe/London
BEGIN:DAYLIGHT
TZOFFSETFROM:+0000
TZOFFSETTO:+0100
TZNAME:BST
DTSTART:19700329T010000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0100
TZOFFSETTO:+0000
TZNAME:GMT
DTSTART:19701025T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
CATEGORIES:Artificial Intelligence Research Group Talks (Comp
 uter Laboratory)
SUMMARY:An Introduction to Federated Learning and its Appl
 ications in Medicine - Blaise Thomson\, Bitfount
DTSTART;TZID=Europe/London:20211130T131500
DTEND;TZID=Europe/London:20211130T141500
UID:TALK162610AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/162610
DESCRIPTION:"Join us on Zoom":https://zoom.us/j/99166955895?pw
 d=SzI0M3pMVEkvNmw3Q0dqNDVRalZvdz09\n\nMachine lear
 ning and artificial intelligence continue to deliv
 er groundbreaking advances in healthcare\, promisi
 ng to transform everything from diagnosis to drug 
 discovery. However academic and clinical researche
 rs alike are currently limited by a lack of access
  to the most relevant datasets for their research 
 and ML model training\, mostly due to legitimate p
 rivacy concerns. The resulting legal and technical
  obstacles usually prevent collaborations entirely
 \, or at best incur months and years of complex le
 gal and technical workarounds in order to enable d
 ata sharing.\n\nFederated learning\, in combinatio
 n with other privacy-preserving ML techniques\, is
  a leading candidate for eliminating this friction
 . It enables researchers to gather insights and tr
 ain models on data which (a) remains held at all t
 imes by the organisation which owns it\, and (b) n
 ever gets exposed in its raw form. It enables data
  owners to share the benefits of their data withou
 t giving up control or privacy. Bitfount is a powe
 rful\, easy-to-use platform for federated AI and d
 ata collaboration\, designed to help researchers m
 aximise the impact of their work. In this seminar\
 , members of the Bitfount team will give a practic
 al introduction to federated learning\, its positi
 on in the array of privacy-preserving ML methods\,
  how it can be combined with other techniques such
  as differential privacy\, secure multi-party comp
 utation\, homomorphic encryption and synthetic dat
 a\, and how these techniques can be applied to sol
 ve real-world problems today both in academia and 
 in industry.\n\n*BIO:*\nBlaise Thomson is the foun
 der and CEO of Bitfount and is a widely-recognised
  AI leader\, researcher and successful entrepreneu
 r. He was the founder and CEO of VocalIQ\, which h
 e sold to Apple in 2015\, subsequently leading the
 ir Cambridge\, UK engineering office and holding t
 he role of Chief Architect for Siri Understanding.
  Blaise holds a PhD in Computer Science from the U
 niversity of Cambridge\, where he was also a Resea
 rch Fellow\, and is an Honorary Fellow at the Camb
 ridge Judge Business School.
LOCATION:Zoom
CONTACT:Mateja Jamnik
END:VEVENT
END:VCALENDAR