Research map / Graph learning / GNN architectures
DRGCN: Dynamic Evolving Initial Residual for Deep Graph Convolutional Networks
Lei Zhang, Xiaodong Yan, Jianshan He, Ruopeng Li, Wei Chu
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 node-personalized dynamic residual weights (ablation) | GCNII · Cora, 2 layers | accuracy: 84.6 -> 85.4 | Improved |
| Add node-personalized dynamic residual weights (ablation) | GCNII · Cora, 64 layers | accuracy: 87.1 -> 86.9 | Hurt |
| Add an evolving block to model layer correlation in residual weights (ablation) | Dynamic-residual GCNII (+Dyn) · Cora, 64 layers | accuracy: 86.9 -> 87.5 | Improved |
| Add data augmentation (ablation) | DRGCN · Cora, 64 layers | accuracy: 87.5 -> 88.1 | Improved |
| Add data augmentation | DRGCN · Cora, 140 training nodes | accuracy: 58.0 -> 85.8 | Improved |
| Use mini-batch instead of full-batch training | DRGAT · Cora | memory: 1.28 -> 0.49 | Improved |
| Use mini-batch instead of full-batch training | DRGAT · Cora | accuracy: 88.20 -> 88.00 | No clear effect |
| Use mini-batch instead of full-batch training | DRGAT · ogbn-arxiv | memory: 7.42 -> 2.27 | Improved |
| Use mini-batch instead of full-batch training | DRGAT · ogbn-arxiv | accuracy: 74.16 -> 74.08 | No clear effect |
| Add GIANT-XRT pretrained features | DRGAT · ogbn-arxiv | test accuracy: 74.16 ± 0.07 -> 76.11 ± 0.09 | Improved |
| Add self-knowledge distillation | DRGAT+GIANT-XRT · ogbn-arxiv | test accuracy: 76.11 ± 0.09 -> 76.33 ± 0.08 | Improved |
Closest papers by meaning
- EvolveGCN: Evolving Graph Convolutional Networks for Dynamic Graphs (AAAI 2020 · similarity 0.92)
- On Provable Benefits of Depth in Training Graph Convolutional Networks (NeurIPS 2021 · similarity 0.90)
- Simple and Deep Graph Convolutional Networks (ICML 2020 · similarity 0.90)
- DeepGCNs: Can GCNs Go As Deep As CNNs? (ICCV 2019 · similarity 0.89)
- Deep Graph Neural Networks via Posteriori-Sampling-based Node-Adaptative Residual Module (NeurIPS 2024 · similarity 0.88)
- Generalizing CNNs to graphs with learnable neighborhood quantization (NeurIPS 2024 · similarity 0.88)
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.
