[论文解读] Gazelle: A Low Latency Framework for Secure Neural Network Inference

Chiraag Juvekar, Vinod Vaikuntanathan+1 more|arXiv (Cornell University)|Jan 16, 2018

Cryptography and Data Security参考文献 23被引用 175

一句话总结

Gazelle 将打包的加法同态加密和混淆电路结合起来，以实现安全的 CNN 推理，其在线延迟和带宽显著低于现有方法。

ABSTRACT

The growing popularity of cloud-based machine learning raises a natural question about the privacy guarantees that can be provided in such a setting. Our work tackles this problem in the context where a client wishes to classify private images using a convolutional neural network (CNN) trained by a server. Our goal is to build efficient protocols whereby the client can acquire the classification result without revealing their input to the server, while guaranteeing the privacy of the server's neural network. To this end, we design Gazelle, a scalable and low-latency system for secure neural network inference, using an intricate combination of homomorphic encryption and traditional two-party computation techniques (such as garbled circuits). Gazelle makes three contributions. First, we design the Gazelle homomorphic encryption library which provides fast algorithms for basic homomorphic operations such as SIMD (single instruction multiple data) addition, SIMD multiplication and ciphertext permutation. Second, we implement the Gazelle homomorphic linear algebra kernels which map neural network layers to optimized homomorphic matrix-vector multiplication and convolution routines. Third, we design optimized encryption switching protocols which seamlessly convert between homomorphic and garbled circuit encodings to enable implementation of complete neural network inference. We evaluate our protocols on benchmark neural networks trained on the MNIST and CIFAR-10 datasets and show that Gazelle outperforms the best existing systems such as MiniONN (ACM CCS 2017) by 20 times and Chameleon (Crypto Eprint 2017/1164) by 30 times in online runtime. Similarly when compared with fully homomorphic approaches like CryptoNets (ICML 2016) we demonstrate three orders of magnitude faster online run-time.

研究动机与目标

• 解决云端 CNN 推理中客户端保持输入私密、服务器保持模型参数私密的隐私问题。
• 设计一个可扩展系统，实现低在线延迟的安全 CNN 推理。
• 利用同态加密和混淆电路的混合来优化线性与非线性计算阶段。

提出的方法

• 介绍 Gazelle 作为三部分系统：Gazelle 同态层，用于快速 SIMD 加法、SIMD 标量乘法和自同构；Gazelle 线性代数内核，用于快速同态矩阵-向量乘法和卷积；Gazelle 网络推理，在混淆电路和同态运算之间切换以实现完整的 CNN 推理。
• Gazelle 线性代数内核，用于快速同态矩阵-向量乘法和卷积；
• Gazelle 网络推理，在完整 CNN 推理中在混淆电路和同态评估之间切换。

实验结果

研究问题

• RQ1 如何将 CNN 推理任务划分，以利用同态加密对线性层的优势和混淆电路对非线性层的优势？
• RQ2 在安全 CNN 推理中，(F)HE 与混淆电路之间的性能权衡是什么，如何在实际中进行优化？
• RQ3 结合 PAHE-GC 的方案是否能实现比现有安全推理系统更低的在线延迟和带宽？

主要发现

• Gazelle 在 CIFAR-10 上比 MiniONN (ACM CCS 2017) 快 20x 的在线运行时间，比 Chameleon (Crypto Eprint 2017/1164) 快 30x。
• Gazelle 在端到端安全推理方面比 CryptoNets (ICML 2016) 的在线运行时间快三个数量级。
• 每次推理的在线带宽降至大约 0.3 GB，CIFAR-10 网络，较 MiniONN 的 6.2 GB。
• Gazelle 在局域网设置下的端到端延迟为 3.6 秒，MiniONN 在 CIFAR-10 网络为 72 秒。
• 系统在 Gazelle 同态层的线性代数基本运算上，接近明文级别性能，相对于明文有 10-20x 的 slowdown。