Skip to Main content Skip to Navigation
Conference papers

SPEED: secure, PrivatE, and efficient deep learning

Arnaud Grivet Sébert 1, * Rafaël Pinot 1, 2, 3 Martin Zuber 4 Cedric Gouy-Pailler 1 Renaud Sirdey 4 
* Corresponding author
1 LI3A - Intelligence Artificielle et Apprentissage Automatique
DM2I - Département Métrologie Instrumentation & Information : DRT/LIST/DM2I
4 LCYL - Laboratoire pour la Confiance des sYstèmes de calcuL
Université Paris-Saclay, DSCIN - Département Systèmes et Circuits Intégrés Numériques : DRT/LIST/DSCIN
Abstract : We introduce a deep learning framework able to deal with strong privacy constraints. Based on collaborative learning, differential privacy and homomorphic encryption, the proposed approach advances state-of-the-art of private deep learning against a wider range of threats, in particular the honest-but-curious server assumption. We address threats from both the aggregation server, the global model and potentially colluding data holders. Building upon distributed differential privacy and a homomorphic argmax operator, our method is specifically designed to maintain low communication loads and efficiency. The proposed method is supported by carefully crafted theoretical results. We provide differential privacy guarantees from the point of view of any entity having access to the final model, including colluding data holders, as a function of the ratio of data holders who kept their noise secret. This makes our method practical to real-life scenarios where data holders do not trust any third party to process their datasets nor the other data holders. Crucially the computational burden of the approach is maintained reasonable, and, to the best of our knowledge, our framework is the first one to be efficient enough to investigate deep learning applications while addressing such a large scope of threats. To assess the practical usability of our framework, experiments have been carried out on image datasets in a classification context. We present numerical results that show that the learning procedure is both accurate and private.
Complete list of metadata
Contributor : Marie-France Robbe Connect in order to contact the contributor
Submitted on : Thursday, July 22, 2021 - 10:48:09 AM
Last modification on : Wednesday, March 16, 2022 - 3:53:59 AM
Long-term archiving on: : Saturday, October 23, 2021 - 6:16:33 PM


 Restricted access
To satisfy the distribution rights of the publisher, the document is embargoed until : 2023-03-23

Please log in to resquest access to the document



Arnaud Grivet Sébert, Rafaël Pinot, Martin Zuber, Cedric Gouy-Pailler, Renaud Sirdey. SPEED: secure, PrivatE, and efficient deep learning. ECML PKDD 2021 - European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, Basque Center for Applied Mathematics, Sep 2021, Bilbao, Spain. pp.675-694, ⟨10.1007/s10994-021-05970-3⟩. ⟨cea-03295491⟩



Record views