TenSEAL: A Library for Encrypted Tensor Operations Using Homomorphic Encryption

TL;DR

TenSEAL is an open-source library enabling privacy-preserving machine learning by performing encrypted tensor operations efficiently with homomorphic encryption, demonstrated on neural networks with promising speed and communication metrics.

Contribution

The paper introduces TenSEAL, a novel library that simplifies integrating homomorphic encryption into machine learning workflows for privacy preservation.

Findings

Encrypted CNN evaluation in under a second

Less than half a megabyte of communication required

Effective benchmarking on MNIST dataset

Abstract

Machine learning algorithms have achieved remarkable results and are widely applied in a variety of domains. These algorithms often rely on sensitive and private data such as medical and financial records. Therefore, it is vital to draw further attention regarding privacy threats and corresponding defensive techniques applied to machine learning models. In this paper, we present TenSEAL, an open-source library for Privacy-Preserving Machine Learning using Homomorphic Encryption that can be easily integrated within popular machine learning frameworks. We benchmark our implementation using MNIST and show that an encrypted convolutional neural network can be evaluated in less than a second, using less than half a megabyte of communication.