Low Latency Privacy Preserving Inference

Abstract

When applying machine learning to sensitive data, one has to find a balance between accuracy, information security, and computational-complexity. Recent studies combined Homomorphic Encryption with neural networks to make inferences while protecting against information leakage. However, these methods are limited by the width and depth of neural networks that can be used (and hence the accuracy) and exhibit high latency even for relatively simple networks. In this study we provide two solutions that address these limitations. In the first solution, we present more than $10\times$ improvement in latency and enable inference on wider networks compared to prior attempts with the same level of security. The improved performance is achieved by novel methods to represent the data during the computation. In the second solution, we apply the method of transfer learning to provide private inference services using deep networks with latency of $\sim 0.16$ seconds. We demonstrate the efficacy of our methods on several computer vision tasks.

Cited by in corpus (19)

• EVA: An Encrypted Vector Arithmetic Language and Compiler for Efficient Homomorphic Computation
• HEAAN Demystified: Accelerating Fully Homomorphic Encryption Through Architecture-centric Analysis and Optimization
• A Study of Face Obfuscation in ImageNet
• Enabling Homomorphically Encrypted Inference for Large DNN Models
• nGraph-HE2: A High-Throughput Framework for Neural Network Inference on Encrypted Data
• F1: A Fast and Programmable Accelerator for Fully Homomorphic Encryption (Extended Version)
• Glyph: Fast and Accurately Training Deep Neural Networks on Encrypted Data
• On Polynomial Approximations for Privacy-Preserving and Verifiable ReLU Networks
• AutoPrivacy: Automated Layer-wise Parameter Selection for Secure Neural Network Inference
• Popcorn: Paillier Meets Compression For Efficient Oblivious Neural Network Inference
• Practical Encrypted Computing for IoT Clients
• Efficient CNN Building Blocks for Encrypted Data
• Efficient Encrypted Inference on Ensembles of Decision Trees
• ENSEI: Efficient Secure Inference via Frequency-Domain Homomorphic Convolution for Privacy-Preserving Visual Recognition
• Enabling Inference Privacy with Adaptive Noise Injection
• BUNET: Blind Medical Image Segmentation Based on Secure UNET
• Application of Homomorphic Encryption in Medical Imaging
• Secure and Differentially Private Bayesian Learning on Distributed Data
• Accelerating 2PC-based ML with Limited Trusted Hardware