Research map / Graph learning / Graph algorithms and clustering
A Scalable Inter-edge Correlation Modeling in CopulaGNN for Link Sign Prediction
Jinkyu Sung, Myunggeum Jee, Joonseok Lee
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 SNEA's two-stage pipeline with Gaussian copula edge correlation model | SNEA backbone · WikiRfa | total training time (sec): 2174.1 -> 18.0 | Improved |
| replace SNEA with CopulaLSP | SNEA backbone · Epinions | total training time (sec): 6574.4 -> 24.7 | Improved |
| replace SNEA with CopulaLSP | SNEA backbone · BitcoinAlpha | Macro F1: 0.669 -> 0.716 | Improved |
| replace SNEA with CopulaLSP | SNEA backbone · BitcoinAlpha | AUC: 0.866 -> 0.864 | No clear effect |
| replace SNEA with CopulaLSP | SNEA backbone · Epinions | GPU memory (GB), training: 7.03 -> 7.20 | Hurt |
| learn Gramian correlation instead of identity (no correlation) (ablation) | CopulaLSP · BitcoinAlpha | Macro F1: 0.665 -> 0.713 | Improved |
| learn Gramian correlation instead of identity (no correlation) (ablation) | CopulaLSP · Epinions | Macro F1: 0.782 -> 0.810 | Improved |
| learn Gramian correlation instead of identity (no correlation) (ablation) | CopulaLSP · BitcoinAlpha, BitcoinOTC, WikiElec, WikiRfa, SlashDot, Epinions | AUC: 0.830 -> 0.865 (BitcoinAlpha); others within 0.001 | No clear effect |
| learn Gramian correlation instead of identity (no correlation) (ablation) | CopulaLSP · WikiElec | epochs to converge: 136.5 -> 204.6 | Hurt |
| learn Gramian correlation instead of identity (no correlation) (ablation) | CopulaLSP · Epinions | epochs to converge: 127.2 -> 58.2 | Improved |
| add Woodbury reformulation for inference (ablation) | CopulaLSP · BitcoinOTC | inference time (sec): 0.64 -> 0.07 | Improved |
| add Woodbury reformulation for inference (ablation) | CopulaLSP · Epinions | inference GPU memory (GB): 1182.54 -> 6.18 | Improved |
Closest papers by meaning
- Self-Explainable Graph Transformer for Link Sign Prediction (AAAI 2025 · similarity 0.89)
- Beyond Node-Centric Modeling: Sketching Signed Networks with Simplicial Complexes (NeurIPS 2025 · similarity 0.89)
- DropEdge not Foolproof: Effective Augmentation Method for Signed Graph Neural Networks (NeurIPS 2024 · similarity 0.88)
- Structure Balance and Gradient Matching-Based Signed Graph Condensation (AAAI 2025 · similarity 0.86)
- Implicit degree bias in the link prediction task (ICML 2025 · similarity 0.86)
- Efficient Graph Matching for Correlated Stochastic Block Models (NeurIPS 2024 · 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.
