Research map / Architectures and efficiency / RNNs and normalization
CryptoGCN: Fast and Scalable Homomorphically Encrypted Graph Convolutional Network Inference
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| replace row-major ciphertext format with AMA format (batch 1) (ablation) | ST-GCN HE inference · NTU-XView | latency: 2962.46s -> 1863.96s (1.59x); 9589.59s -> 4850.93s (1.97x) | Helped |
| increase batch size from 1 to 16 with AMA vs row-major (ablation) | AMA format · NTU-XView 64-STGCN-3 | speedup: 1.59x -> 3.10x | Helped |
| prune one activation layer (1 AP) (ablation) | CryptoGCN · NTU-XView | latency: 4273.89s -> 1863.95s (64-STGCN-3) | Helped |
| prune one activation layer (1 AP) (ablation) | CryptoGCN · NTU-XView | accuracy: 74.25 -> 73.12 (64-STGCN-3); 75.31 -> 73.78 (128) | Hurt |
| prune a second activation layer (2 AP vs 1 AP) (ablation) | CryptoGCN · NTU-XView | accuracy: 73.12 -> 70.21 (64-STGCN-3) | Hurt |
| prune a second activation layer (2 AP vs 1 AP) (ablation) | CryptoGCN · NTU-XView | latency: 1863.95s -> 1856.36s | No clear effect |
| AMA format vs CHET/Fast-HEAR row-major packing | 64-STGCN-3 · NTU-XView | homomorphic operation count: 1.86M -> 589K total HOC (batch 1, vs Fast-HEAR) | Helped |
Closest papers by meaning
- FicGCN: Unveiling the Homomorphic Encryption Efficiency from Irregular Graph Convolutional Networks (ICML 2025 · similarity 0.93)
- DCT-CryptoNets: Scaling Private Inference in the Frequency Domain (ICLR 2025 · similarity 0.88)
- PrivCirNet: Efficient Private Inference via Block Circulant Transformation (NeurIPS 2024 · similarity 0.86)
- GRAIN: Exact Graph Reconstruction from Gradients (ICLR 2025 · similarity 0.85)
- Converting Transformers to Polynomial Form for Secure Inference Over Homomorphic Encryption (ICML 2024 · similarity 0.85)
- CipherPrune: Efficient and Scalable Private Transformer Inference (ICLR 2025 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.