Efficient privacy-preserving inference for convolutional neural networks

TL;DR

This paper introduces a method to improve privacy-preserving CNN inference by optimizing tensor representations, significantly reducing computational costs on encrypted data for datasets like MNIST and CIFAR-10.

Contribution

It proposes novel tensor representation techniques that halve the number of homomorphic encryption operations needed during CNN inference.

Findings

Over two-fold reduction in HE operations compared to CryptoNets

Effective application on MNIST and CIFAR-10 datasets

Enhanced efficiency in privacy-preserving CNN inference

Abstract

The processing of sensitive user data using deep learning models is an area that has gained recent traction. Existing work has leveraged homomorphic encryption (HE) schemes to enable computation on encrypted data. An early work was CryptoNets, which takes 250 seconds for one MNIST inference. The main limitation of such approaches is that of the expensive FFT-like operations required to perform operations on HE-encrypted ciphertext. Others have proposed the use of model pruning and efficient data representations to reduce the number of HE operations required. We focus on improving upon existing work by proposing changes to the representations of intermediate tensors during CNN inference. We construct and evaluate private CNNs on the MNIST and CIFAR-10 datasets, and achieve over a two-fold reduction in the number of operations used for inferences of the CryptoNets architecture.