Towards the AlexNet Moment for Homomorphic Encryption: HCNN, the First Homomorphic CNN on Encrypted Data With GPUs

• Published in: IEEE transactions on emerging topics in computing Vol. 9; no. 3; pp. 1330 - 1343
• Main Authors: Al Badawi, Ahmad, Jin, Chao, Lin, Jie, Mun, Chan Fook, Jie, Sim Jun, Tan, Benjamin Hong Meng, Nan, Xiao, Aung, Khin Mi Mi, Chandrasekhar, Vijay Ramaseshan
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.07.2021; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Artificial neural networks; Cloud computing; Computational modeling; Deep learning; Encryption; GPUs; Graphics processing units; homomorphic encryption; Image classification; implementation; Machine learning; Privacy; privacy-preserving technologies; Servers; Training
• ISSN: 2168-6750; 2168-6750
• DOI: 10.1109/TETC.2020.3014636

Deep Learning as a Service (DLaaS) stands as a promising solution for cloud-based inference applications. In this setting, the cloud has a pre-learned model whereas the user has samples on which she wants to run the model. The biggest concern with DLaaS is the user privacy if the input samples are sensitive data. We provide here an efficient privacy-preserving system by employing high-end technologies such as Fully Homomorphic Encryption (FHE), Convolutional Neural Networks (CNNs) and Graphics Processing Units (GPUs). FHE, with its widely-known feature of computing on encrypted data, empowers a wide range of privacy-concerned applications. This comes at high cost as it requires enormous computing power. In this article, we show how to accelerate the performance of running CNNs on encrypted data with GPUs. We evaluated two CNNs to classify homomorphically the MNIST and CIFAR-10 datasets. Our solution achieved sufficient security level (<inline-formula><tex-math notation="LaTeX">> 80</tex-math> <mml:math><mml:mrow><mml:mo>></mml:mo><mml:mn>80</mml:mn></mml:mrow></mml:math><inline-graphic xlink:href="qaisarahmadalbadawi-ieq1-3014636.gif"/> </inline-formula> bit) and reasonable classification accuracy (99) and (77.55 percent) for MNIST and CIFAR-10, respectively. In terms of latency, we could classify an image in 5.16 seconds and 304.43 seconds for MNIST and CIFAR-10, respectively. Our system can also classify a batch of images (<inline-formula><tex-math notation="LaTeX">></tex-math> <mml:math><mml:mo>></mml:mo></mml:math><inline-graphic xlink:href="qaisarahmadalbadawi-ieq2-3014636.gif"/> </inline-formula> 8,000) without extra overhead.