TenSEAL: A Library for Encrypted Tensor Operations Using Homomorphic Encryption
This addresses privacy concerns for users handling sensitive data like medical and financial records in machine learning applications, though it is incremental as it builds on existing homomorphic encryption techniques.
The authors tackled the problem of privacy threats in machine learning by developing TenSEAL, an open-source library for encrypted tensor operations using homomorphic encryption, which enables evaluation of an encrypted convolutional neural network on MNIST in under a second with less than half a megabyte of communication.
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.