Research map / Graph learning / GNN architectures
Co-Embedding of Nodes and Edges With Graph Neural Networks
Xiaodong Jiang, Ronghang Zhu, Pengsheng Ji, Sheng Li
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 |
|---|---|---|---|
| add edge-node switching convolution (CensNet instead of GCN) | GCN · Tox21 test (90% train) | AUC: 0.75 -> 0.79 | Improved |
| add edge-node switching convolution (CensNet instead of GCN) | GCN · Lipophilicity test (90% train) | RMSE: 0.97 -> 0.83 | Improved |
| add edge-node switching convolution (CensNet instead of GCN) | GCN · Lipophilicity test (60% train) | RMSE: 1.06 -> 0.97 | Improved |
| add edge-node switching convolution (CensNet instead of GCN) | GCN · Cora node classification (3% labels) | accuracy (%): 74.0 -> 79.4 | Improved |
| add edge-node switching convolution (CensNet instead of GCN) | GCN · Citeseer node classification (0.3% labels) | accuracy (%): 39.2 -> 49.4 | Improved |
| add edge-node switching convolution (CensNet instead of AdaLNet) | AdaLNet · Cora node classification (0.5% labels) | accuracy (%): 60.8 -> 57.7 | Hurt |
| add edge-node switching convolution (CensNet instead of LNet) | LNet · PubMed node classification (0.1% labels) | accuracy (%): 73.4 -> 69.9 | Hurt |
| add edge-node switching to the VAE encoder (CensNet-VAE instead of VGAE) | VGAE · Pubmed link prediction | AUC (%): 94.4 -> 95.5 | Improved |
| add edge-node switching to the VAE encoder (CensNet-VAE instead of VGAE) | VGAE · Citeseer link prediction | AUC (%): 90.8 -> 90.6 | No clear effect |
| add more edge layers to CensNet-VAE (ablation) | CensNet-VAE · link prediction (Cora, Citeseer, Pubmed) | AUC | No clear effect |
| add node-edge switching operation | GCN · Cora node classification | time per epoch: about 40% slower | Hurt |
| use mini-batch training with proportional sampling from train/val/test (ablation) | CensNet · node classification | accuracy | No clear effect |
Closest papers by meaning
- Graph Neural Networks Powered by Encoder Embedding for Improved Node Learning (TPAMI 2026 · similarity 0.91)
- EdgeNets: Edge Varying Graph Neural Networks (TPAMI 2021 · similarity 0.90)
- Iterative Deep Graph Learning for Graph Neural Networks: Better and Robust Node Embeddings (NeurIPS 2020 · similarity 0.90)
- Fisher Information Embedding for Node and Graph Learning (ICML 2023 · similarity 0.90)
- L2-GCN: Layer-Wise and Learned Efficient Training of Graph Convolutional Networks (CVPR 2020 · similarity 0.90)
- Towards characterizing the value of edge embeddings in Graph Neural Networks (ICML 2025 · similarity 0.89)
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.
