TAPAS: Tricks to Accelerate (encrypted) Prediction As a Service

TL;DR

This paper introduces TAPAS, a set of techniques that significantly accelerate encrypted prediction services by combining neural network binarization, sparsification, and parallel computation to reduce the computational overhead of homomorphic encryption.

Contribution

It presents novel methods to speed up encrypted neural network inference by integrating binarization, sparsification, and parallelization techniques within the homomorphic encryption framework.

Findings

Achieved substantial reduction in computation time for encrypted predictions.

Demonstrated the effectiveness of binarization and sparsification in encrypted neural networks.

Provided scalable solutions for privacy-preserving prediction services.

Abstract

Machine learning methods are widely used for a variety of prediction problems. \emph{Prediction as a service} is a paradigm in which service providers with technological expertise and computational resources may perform predictions for clients. However, data privacy severely restricts the applicability of such services, unless measures to keep client data private (even from the service provider) are designed. Equally important is to minimize the amount of computation and communication required between client and server. Fully homomorphic encryption offers a possible way out, whereby clients may encrypt their data, and on which the server may perform arithmetic computations. The main drawback of using fully homomorphic encryption is the amount of time required to evaluate large machine learning models on encrypted data. We combine ideas from the machine learning literature, particularly work on binarization and sparsification of neural networks, together with algorithmic tools to speed-up and parallelize computation using encrypted data.