Research map / Graph learning / GNN architectures
An Annotated Graph Model with Differential Degree Heterogeneity for Directed Networks
Stefan Stein, Chenlei Leng
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 |
|---|---|---|---|
| select lambda by BIC instead of theory-based heuristic (ablation) | Penalized MLE for annotated graph model · Simulation (small n) | estimation error of alpha, beta, mu | Improved |
| select lambda by BIC instead of theory-based heuristic (ablation) | Penalized MLE · Simulation | estimation error of gamma | No clear effect |
| select lambda by theory-based heuristic instead of BIC (ablation) | Penalized MLE · Simulation (growing n) | probability of correct model selection | Improved |
| select lambda by theory-based heuristic instead of BIC (ablation) | Penalized MLE · Simulation (small n) | false negatives | Hurt |
| select lambda by BIC (ablation) | Penalized MLE · Simulation (large n) | probability of correct model selection | Hurt |
| increase network size n from 200 to 800 (ablation) | Penalized MLE confidence interval for gamma_0,1 · Simulation | median CI length (heuristic): 0.265 -> 0.075 | Improved |
| use BIC versus heuristic lambda for gamma_0,1 inference (ablation) | Penalized MLE · Simulation | coverage at n=200: 0.952 -> 0.962 | No clear effect |
| use sparse penalized model on all 71 nodes instead of dense MLE model on 63 nodes | Annotated graph model · Lazega lawyer network | standard errors: 0.10 vs 0.16 (Same status SE) | Improved |
| use annotated graph model with sparse alpha/beta instead of Yan et al. model | Annotated graph model vs Yan et al. · Lazega lawyer network | Same law school coefficient sign: 0.29 vs -0.48 | Mixed |
| use annotated graph model instead of Yan et al. model | Sina Weibo network · Sina Weibo network | coefficient signs without bias correction: -0.127 vs -0.391 (posts) | Improved |
| vanilla 10-fold cross-validation link prediction | Annotated graph model · Lazega lawyer network link prediction | AUC: high 80% range | Improved |
Closest papers by meaning
- Understanding Heterophily for Graph Neural Networks (ICML 2024 · similarity 0.87)
- Uncertainty Estimation for Heterophilic Graphs Through the Lens of Information Theory (ICML 2025 · similarity 0.87)
- Graph Structure Inference with BAM: Neural Dependency Processing via Bilinear Attention (NeurIPS 2024 · similarity 0.86)
- On the Role of Edge Dependency in Graph Generative Models (ICML 2024 · similarity 0.86)
- A Generic Family of Graphical Models: Diversity, Efficiency, and Heterogeneity (ICML 2025 · similarity 0.86)
- When Edge Independence Fails: Joint Graph Diffusion with Latent Sociability Priors (NeurIPS 2026 · similarity 0.86)
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.
