TGHE: Template-based Graph Homomorphic Encryption for Privacy-Preserving GNN Inference in Edge-Cloud Systems
arXiv:2606.26664v1 Announce Type: cross Abstract: Existing homomorphic encryption (HE)-based GNN systems adopt a graph-centric paradigm that couples per-query cost to global graph size, limiting evaluations to at most ~20k nodes and making them incompatible with dynamic, large-scale financial graphs. We propose TGHE (Template-based Graph Homomorphic Encryption), an ego-centric framework that resolves this by exploiting a template phenomenon: local computation trees in transaction graphs converge into a small set of structural shapes. TGHE canonicalizes ego-graphs at the edge and packs structur
The increasing demand for privacy-preserving AI, especially in sensitive financial applications and edge computing, necessitates more scalable and efficient homomorphic encryption solutions.
This breakthrough addresses a critical scalability bottleneck in privacy-preserving GNN inference, potentially enabling a wider adoption of secure AI in financial and other data-sensitive industries.
Homomorphic encryption-based GNN systems can now efficiently process larger and more dynamic graphs, moving beyond previous limitations tied to global graph size.
- · Financial institutions
- · Edge computing providers
- · Privacy-preserving AI developers
- · Graph Neural Network applications
- · Traditional privacy-invasive GNN inference methods
Wider deployment of secure GNNs for fraud detection and financial analysis becomes technically feasible.
Reduced regulatory hurdles and increased public trust for AI systems handling sensitive personal or financial data due to enhanced privacy guarantees.
The development of a new ecosystem of privacy-preserving AI services that operate directly on encrypted, dynamic, and large-scale datasets.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.AI