Research map / Graph learning / GNN architectures
coVariance Neural Networks
Saurabh Sihag, Gonzalo Mateos, Corey McMillan, Alejandro Ribeiro
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 |
|---|---|---|---|
| VNN instead of PCA-LR | age regression · FTDC100 | MAE: 5.19 -> 4.31 | Improved |
| VNN instead of PCA-LR | age regression · FTDC300 | MAE: 5.97 -> 4.45 | Improved |
| VNN instead of PCA-LR | age regression · FTDC500 | MAE: 5.31 -> 4.74 | Improved |
| VNN instead of PCA-rbf | age regression · FTDC100 | MAE: 4.49 -> 4.31 | Improved |
| VNN instead of PCA-rbf | age regression · FTDC300 | MAE: 4.64 -> 4.45 | Improved |
| VNN instead of PCA-rbf | age regression · FTDC500 | MAE: 4.38 -> 4.74 | Hurt |
| VNN instead of PCA-rbf | age regression · FTDC300 | correlation: 0.61 -> 0.62 | No clear effect |
| perturb sample covariance matrix (add/remove samples) | PCA-regression · ABC test set | MAE | Hurt |
| perturb sample covariance matrix | VNN · ABC test set | MAE | No clear effect |
| reduce sample size below 50 for covariance (ablation) | VNN · ABC test set | MAE | Hurt |
| perturb covariance matrix with rbf regression | PCA-rbf · FTDC100/300 | MAE | No clear effect |
| transfer VNN across FTDC resolutions without retraining | VNN · FTDC100/300/500 | MAE and correlation | No clear effect |
Closest papers by meaning
- Neural Tangent Kernels Motivate Cross-Covariance Graphs in Neural Networks (ICML 2024 · similarity 0.88)
- On the Stability of Graph Convolutional Neural Networks: A Probabilistic Perspective (NeurIPS 2025 · similarity 0.87)
- Graph Neural Networks Are More Than Filters: Revisiting and Benchmarking from A Spectral Perspective (ICLR 2025 · similarity 0.87)
- A Manifold Perspective on the Statistical Generalization of Graph Neural Networks (ICML 2025 · similarity 0.86)
- Improving GNN Calibration with Discriminative Ability: Insights and Strategies (AAAI 2024 · similarity 0.86)
- Exploring Consistency in Graph Representations: from Graph Kernels to Graph Neural Networks (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.
