BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:CIS - "Get to know your neighbors" Seminar series - Prof. Jean-Pie
 rre Hubaux
DTSTART:20211011T151500
DTEND:20211011T161500
DTSTAMP:20260916T064128Z
UID:b96f91749061145bbe7ee77e9f7c88efda3e2593814a30c035c243a4
CATEGORIES:Conferences - Seminars
DESCRIPTION:Prof Jean-Pierre Hubaux\n \nTitle: Secure and Privacy-Conscio
 us Federated Analytics\nJean-Pierre Hubaux\, Head of Laboratory for Data S
 ecurity (LDS)\, EPFL\n\nAbstract: In this talk\, we address the problem o
 f privacy-preserving training and evaluation of neural networks in an N-pa
 rty\, federated learning setting. We propose a novel system\, POSEIDON\, t
 he first of its kind in the regime of privacy-preserving neural network tr
 aining. It employs multiparty lattice-based cryptography to preserve the c
 onfidentiality of the training data\, the model\, and the evaluation data\
 , under a passive-adversary model and collusions between up to N−1 parti
 es. It relies on homomorphic encryption and secure multi-party computation
 . We also introduce arbitrary linear transformations within the cryptograp
 hic bootstrapping operation\, optimizing the costly cryptographic computat
 ions over the parties. We also mention Lattigo\, our quantum-resistant ope
 n-source cryptographic library on which POSEIDON is based. Our experimenta
 l results show that POSEIDON achieves accuracy similar to centralized (or 
 decentralized) non-private approaches and that its computation and communi
 cation overhead scales linearly with the number of parties. Furthermore\, 
 we explain how we are using these techniques for the federated analysis o
 f medical data\, in particular for genome-wide association studies. Final
 ly\, we mention our joint work with lawyers showing GDPR compliance\,  an
 d address the creation of start-up Tune Insight.\n\nBio: Jean-Pierre Hubau
 x is a full professor at EPFL and head of the Laboratory for Data Securi
 ty. Through his research\, he contributes to laying the foundations and de
 veloping the tools for protecting privacy in today’s hyper-connected w
 orld. He has pioneered the areas of privacy and security in mobile/wirele
 ss networks and in personalized health.\nHe is the academic director of 
 the Center for Digital Trust (C4DT). He leads the  Data Protection in P
 ersonalized Health (DPPH) project funded by the ETH Council. He is a Fell
 ow of both IEEE (2008) and ACM (2010). Recent awards: three of his papers 
 obtained distinctions at the IEEE Symposium on Security and Privacy in 20
 15\, 2018 and 2021. He is among the most cited researchers in privacy pro
 tection and in information security. More about him here.\n\n \nThe CIS 
 seminar will take place live on Zoom: https://epfl.zoom.us/j/65412917575 \
 n\nPlease connect to your zoom account using your "@epfl.ch" address\, as 
 this live event is only open to the EPFL community\nMonday\, October 11th\
 , 2021 from 3:15 to 4:15 pm\nNB: Video recordings of the seminars will be 
 made available on our website and published on our social media pages
LOCATION:INF 328 https://plan.epfl.ch/?room==INF%20328 https://epfl.zoom.u
 s/j/65412917575
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR