BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Secure Multi-Party Linear Regression on High-Dimensional Data - Bo
 rja Balle (Amazon)
DTSTART:20171019T140000Z
DTEND:20171019T150000Z
UID:TALK82901@talks.cam.ac.uk
CONTACT:Liang Wang
DESCRIPTION:The goal of secure multi-pary computation (MPC) is to facilita
 te the evaluation of functionalities that depend on the private inputs of 
 several distrusting parties in a privacy preserving manner. I will start m
 y talk by discussing potential applications of secure MPC to machine learn
 ing and the relation between MPC and other well-known privacy frameworks l
 ike differential privacy. Then I will present our recent work on secure MP
 C protocols for linear regression on distributed databases. By combining s
 everal tools from the MPC literature we obtain scalable solutions that can
  solve problems with millions of records and hundreds of features in a mat
 ter of minutes. Some crucial implementation details will be discussed\, in
 cluding the role of fixed-point arithmetic and a robust conjugate gradient
  descent solver for private linear systems. An implementation of our proto
 cols based on the Obliv-C framework is available as open source.\n \nBio: 
 Borja Balle is currently a Machine Learning Scientist at Amazon Research C
 ambridge. Before joining Amazon\, Borja was a lecturer at Lancaster Univer
 sity (2015-2017) and a postdoctoral fellow at McGill University (2013-2015
 ). His main research interest is in privacy-preserving machine learning\, 
 including the use of differential privacy and multi-party computation in d
 istributed learning problems\, and the foundations of privacy-aware data s
 cience. More info at https://borjaballe.github.io
LOCATION:FW26\, Computer Laboratory\, William Gates Building
END:VEVENT
END:VCALENDAR