AriaNN: Low-Interaction Privacy-Preserving Deep Learning via Function Secret Sharing
Théo Ryffel, Pierre Tholoniat, David Pointcheval, Francis Bach
Abstract
We propose AriaNN, a low-interaction privacy-preserving framework for private neural network training and inference on sensitive data.
Topics & Concepts
Computer scienceSecret sharingBottleneckServerCryptographySecure multi-party computationProtocol (science)Cryptographic primitiveArtificial neural networkTheoretical computer scienceInferenceCryptographic protocolArtificial intelligenceDistributed computingComputer securityComputer networkEmbedded systemMedicinePathologyAlternative medicinePrivacy-Preserving Technologies in DataCryptography and Data SecurityAdversarial Robustness in Machine Learning