FicGCN: Unveiling the Homomorphic Encryption Efficiency from Irregular Graph Convolutional Networks

TL;DR

FicGCN is a novel homomorphic encryption framework optimized for sparse graph convolutional networks, significantly improving privacy-preserving GCN computation efficiency by reducing rotation overhead and balancing operations.

Contribution

The paper introduces FicGCN, a new HE-based framework that leverages GCN sparsity and region-based data reordering to enhance privacy-preserving graph learning performance.

Findings

Achieved up to 4.10x performance improvement over existing methods.

Effectively reduces rotation overhead in HE-based GCN computations.

Demonstrated superior performance across multiple datasets.

Abstract

Graph Convolutional Neural Networks (GCNs) have gained widespread popularity in various fields like personal healthcare and financial systems, due to their remarkable performance. Despite the growing demand for cloud-based GCN services, privacy concerns over sensitive graph data remain significant. Homomorphic Encryption (HE) facilitates Privacy-Preserving Machine Learning (PPML) by allowing computations to be performed on encrypted data. However, HE introduces substantial computational overhead, particularly for GCN operations that require rotations and multiplications in matrix products. The sparsity of GCNs offers significant performance potential, but their irregularity introduces additional operations that reduce practical gains. In this paper, we propose FicGCN, a HE-based framework specifically designed to harness the sparse characteristics of GCNs and strike a globally optimal balance between aggregation and combination operations. FicGCN employs a latency-aware packing scheme, a Sparse Intra-Ciphertext Aggregation (SpIntra-CA) method to minimize rotation overhead, and a region-based data reordering driven by local adjacency structure. We evaluated FicGCN on several popular datasets, and the results show that FicGCN achieved the best performance across all tested datasets, with up to a 4.10x improvement over the latest design.