Fully Homomorphic Encryption (FHE) has the potential to substantially improve privacy and security by enabling computation directly on encrypted data. This is especially true with deep learning, as today, many popular user services are powered by neural networks in the cloud. One of the major challenges facing wide-scale deployment of FHE-secured neural inference is effectively mapping these networks to FHE primitives. FHE poses many programming challenges including packing large vectors, automatically managing noise via bootstrapping, and translating arbitrary and general-purpose programs to the limited instruction set provided by FHE. These challenges make building large FHE neural networks intractable using the tools available today. In this paper we address these challenges with Orion, a fully-automated framework for private neural inference in FHE. Orion accepts deep neural networks written in PyTorch and translates them into efficient FHE programs. We achieve this by proposing a novel single-shot multiplexed packing strategy for arbitrary convolutions and through a new, efficient technique to automate bootstrap placement. We evaluate Orion on common benchmarks used by the FHE deep learning community and outperform state-of-the-art by $2.38 \times$ on ResNet-20, the largest network they report. Orion extends naturally to larger networks. We demonstrate this by evaluating ResNet-50 on ImageNet and present the first high-resolution homomorphic object detection experiments using a YOLO-v1 model with 139 million parameters. Finally, we open-source our framework Orion at the following repository: https://github.com/baahl-nyu/orion

翻译：全同态加密（FHE）能够直接在加密数据上进行计算，有望显著提升隐私与安全性。这一特性在深度学习领域尤为重要，因为当今许多主流用户服务都依赖于云端神经网络。大规模部署基于FHE的神经推理面临的主要挑战之一，是如何高效地将这些网络映射到FHE原语。FHE带来了诸多编程难题，包括大规模向量打包、通过自举自动管理噪声，以及将任意通用程序转换为FHE有限指令集。这些挑战使得利用现有工具构建大型FHE神经网络变得异常困难。本文通过提出Orion框架来解决这些问题，该框架是实现FHE私有神经推理的全自动化系统。Orion能够接收以PyTorch编写的深度神经网络，并将其转换为高效的FHE程序。我们通过以下两项创新实现这一目标：针对任意卷积的新型单次复用打包策略，以及自动化自举布局的高效新技术。我们在FHE深度学习社区常用基准测试上评估Orion，在ResNet-20（该领域已报道的最大网络）上性能超越现有最佳方案$2.38 \times$。Orion可自然扩展至更大规模网络，我们通过在ImageNet上评估ResNet-50证明了这一点，并首次使用包含1.39亿参数的YOLO-v1模型实现了高分辨率同态目标检测实验。最后，我们在以下代码库开源Orion框架：https://github.com/baahl-nyu/orion