Enabling Homomorphically Encrypted Inference for Large DNN Models
arXiv:2103.16139 · doi:10.1109/TC.2021.3076123
Abstract
The proliferation of machine learning services in the last few years has raised data privacy concerns. Homomorphic encryption (HE) enables inference using encrypted data but it incurs 100x-10,000x memory and runtime overheads. Secure deep neural network (DNN) inference using HE is currently limited by computing and memory resources, with frameworks requiring hundreds of gigabytes of DRAM to evaluate small models. To overcome these limitations, in this paper we explore the feasibility of leveraging hybrid memory systems comprised of DRAM and persistent memory. In particular, we explore the recently-released Intel Optane PMem technology and the Intel HE-Transformer nGraph to run large neural networks such as MobileNetV2 (in its largest variant) and ResNet-50 for the first time in the literature. We present an in-depth analysis of the efficiency of the executions with different hardware and software configurations. Our results conclude that DNN inference using HE incurs on friendly access patterns for this memory configuration, yielding efficient executions.
Manuscript accepted for publication in IEEE Transactions on Computers
References in corpus (10)
- MobileNetV2: Inverted Residuals and Linear Bottlenecks
- A generic framework for privacy preserving deep learning
- CryptoDL: Deep Neural Networks over Encrypted Data
- XONN: XNOR-based Oblivious Deep Neural Network Inference
- Faster CryptoNets: Leveraging Sparsity for Real-World Encrypted Inference
- Intel nGraph: An Intermediate Representation, Compiler, and Executor for Deep Learning
- Low Latency Privacy Preserving Inference
- Private Machine Learning in TensorFlow using Secure Computation
- SEALion: a Framework for Neural Network Inference on Encrypted Data
- nGraph-HE2: A High-Throughput Framework for Neural Network Inference on Encrypted Data