FastFHE: Packing-Scalable and Depthwise-Separable CNN Inference Over FHE

TL;DR

FastFHE introduces a scalable, depthwise-separable CNN inference method over fully homomorphic encryption, significantly reducing latency and storage costs while maintaining high accuracy for secure encrypted deep learning applications.

Contribution

The paper presents a novel FastFHE framework with a new ciphertext packing scheme, depthwise-separable convolutions, BN fusion, and polynomial activation approximation to improve encrypted CNN inference.

Findings

Reduced inference latency and storage costs.

Maintained high accuracy with encrypted CNN models.

Validated efficiency through comprehensive experiments.

Abstract

The deep learning (DL) has been penetrating daily life in many domains, how to keep the DL model inference secure and sample privacy in an encrypted environment has become an urgent and increasingly important issue for various security-critical applications. To date, several approaches have been proposed based on the Residue Number System variant of the Cheon-Kim-Kim-Song (RNS-CKKS) scheme. However, they all suffer from high latency, which severely limits the applications in real-world tasks. Currently, the research on encrypted inference in deep CNNs confronts three main bottlenecks: i) the time and storage costs of convolution calculation; ii) the time overhead of huge bootstrapping operations; and iii) the consumption of circuit multiplication depth. Towards these three challenges, we in this paper propose an efficient and effective mechanism FastFHE to accelerate the model inference while simultaneously retaining high inference accuracy over fully homomorphic encryption. Concretely, our work elaborates four unique novelties. First, we propose a new scalable ciphertext data-packing scheme to save the time and storage consumptions. Second, we work out a depthwise-separable convolution fashion to degrade the computation load of convolution calculation. Third, we figure out a BN dot-product fusion matrix to merge the ciphertext convolutional layer with the batch-normalization layer without incurring extra multiplicative depth. Last but not least, we adopt the low-degree Legendre polynomial to approximate the nonlinear smooth activation function SiLU under the guarantee of tiny accuracy error before and after encrypted inference. Finally, we execute multi-facet experiments to verify the efficiency and effectiveness of our proposed approach.