arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2307.04077cs.CRcs.LG
收藏

迈向快速且可扩展的隐私推理

Towards Fast and Scalable Private Inference

New York University(纽约大学)

Jianqiao Mo, Karthik Garimella, Negar Neda, Austin Ebel, Brandon Reagen

AI总结:

本文针对隐私保护计算开销过高的问题,以神经网络隐私推理为应用,分析同态加密、混淆电路等技术开销并提出HAAC和RPU两种加速方案。

摘要:

隐私与安全已迅速成为首要的设计约束。用户如今要求对其数据谁能查看(机密性)以及如何使用(控制权)获得更强的保护。在此方面,现有的安全密码技术存在不足:它们在数据存储或通信时能够保护数据,但在计算时必须解密。幸运的是,存在一种新的计算范式,我们称之为隐私保护计算(PPC)。新兴的PPC技术可用于安全的外包计算,或使两方在不泄露任何一方用户秘密数据的情况下完成计算。尽管PPC在数字时代革新用户保护方面具有巨大潜力,但由于极高的计算、通信和存储开销,其实际落地一直受到限制。本文回顾了近期以神经网络中的隐私推理(PI)作为驱动应用来解决各类PPC开销的工作。首先,介绍了问题背景以及多种技术,包括同态加密(HE)、秘密共享(SS)、混淆电路(GC)和不经意传输(OT)。接着,阐述了这些技术在用于实现PI时的开销特征。该特征分析表明同时需要GC加速器和HE加速器。随后提出了两种解决方案:用于加速GC的HAAC和用于加速HE的RPU。最后,展示了结果与效果,并讨论了为克服PI剩余开销所需的未来工作。

英文摘要:

Privacy and security have rapidly emerged as first order design constraints. Users now demand more protection over who can see their data (confidentiality) as well as how it is used (control). Here, existing cryptographic techniques for security fall short: they secure data when stored or communicated but must decrypt it for computation. Fortunately, a new paradigm of computing exists, which we refer to as privacy-preserving computation (PPC). Emerging PPC technologies can be leveraged for secure outsourced computation or to enable two parties to compute without revealing either users' secret data. Despite their phenomenal potential to revolutionize user protection in the digital age, the realization has been limited due to exorbitant computational, communication, and storage overheads. This paper reviews recent efforts on addressing various PPC overheads using private inference (PI) in neural network as a motivating application. First, the problem and various technologies, including homomorphic encryption (HE), secret sharing (SS), garbled circuits (GCs), and oblivious transfer (OT), are introduced. Next, a characterization of their overheads when used to implement PI is covered. The characterization motivates the need for both GCs and HE accelerators. Then two solutions are presented: HAAC for accelerating GCs and RPU for accelerating HE. To conclude, results and effects are shown with a discussion on what future work is needed to overcome the remaining overheads of PI.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑