POSEIDON: Privacy-Preserving Federated Neural Network Learning

TL;DR

POSEIDON is a pioneering system that enables privacy-preserving neural network training in federated learning using lattice-based cryptography, achieving accuracy comparable to non-private methods with scalable efficiency.

Contribution

It introduces the first privacy-preserving federated neural network training system employing lattice cryptography, with optimized secure backpropagation and scalable performance.

Findings

Achieves MNIST accuracy similar to non-private training.

Scales linearly in computation and communication with number of parties.

Trains a 3-layer network on 60K samples in under 2 hours.

Abstract

In this paper, we address the problem of privacy-preserving training and evaluation of neural networks in an $N$-party, federated learning setting. We propose a novel system, POSEIDON, the first of its kind in the regime of privacy-preserving neural network training. It employs multiparty lattice-based cryptography to preserve the confidentiality of the training data, the model, and the evaluation data, under a passive-adversary model and collusions between up to $N-1$ parties. To efficiently execute the secure backpropagation algorithm for training neural networks, we provide a generic packing approach that enables Single Instruction, Multiple Data (SIMD) operations on encrypted data. We also introduce arbitrary linear transformations within the cryptographic bootstrapping operation, optimizing the costly cryptographic computations over the parties, and we define a constrained optimization problem for choosing the cryptographic parameters. Our experimental results show that POSEIDON achieves accuracy similar to centralized or decentralized non-private approaches and that its computation and communication overhead scales linearly with the number of parties. POSEIDON trains a 3-layer neural network on the MNIST dataset with 784 features and 60K samples distributed among 10 parties in less than 2 hours.