Faster CryptoNets: Leveraging Sparsity for Real-World Encrypted Inference

TL;DR

This paper introduces Faster CryptoNets, a neural network inference method that uses sparsity, pruning, and quantization to significantly accelerate encrypted computations while maintaining accuracy, enhancing privacy-preserving machine learning.

Contribution

It develops a novel sparse encoding approach and optimal activation function approximation to improve encrypted inference efficiency, incorporating privacy-safe training techniques for real-world datasets.

Findings

Achieves significant speedup over previous encrypted inference methods

Maintains competitive accuracy with reduced computational overhead

Leverages transfer learning and differential privacy for practical applications

Abstract

Homomorphic encryption enables arbitrary computation over data while it remains encrypted. This privacy-preserving feature is attractive for machine learning, but requires significant computational time due to the large overhead of the encryption scheme. We present Faster CryptoNets, a method for efficient encrypted inference using neural networks. We develop a pruning and quantization approach that leverages sparse representations in the underlying cryptosystem to accelerate inference. We derive an optimal approximation for popular activation functions that achieves maximally-sparse encodings and minimizes approximation error. We also show how privacy-safe training techniques can be used to reduce the overhead of encrypted inference for real-world datasets by leveraging transfer learning and differential privacy. Our experiments show that our method maintains competitive accuracy and achieves a significant speedup over previous methods. This work increases the viability of deep learning systems that use homomorphic encryption to protect user privacy.