Research map / Graph learning / GNN architectures
Graph Neural Networks with Adaptive Residual
Xiaorui Liu, Jiayuan Ding, Wei Jin, Han Xu, Yao Ma, Zitao Liu, Jiliang Tang
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 residual connections to message passing (APPNP, GCNII, GCN) (ablation) | GNNs with abnormal node features · Cora/CiteSeer/PubMed with 10% Gaussian-noise nodes | accuracy on abnormal nodes | Hurt |
| add residual connections to message passing (ablation) | GNNs with abnormal node features · Cora/CiteSeer/PubMed | accuracy on normal nodes with more layers | Improved |
| increase number of feature aggregation layers without residual (ablation) | GNNs · Cora/CiteSeer/PubMed | accuracy | Mixed |
| use AirGNN adaptive residual instead of GAT/GCN/GCNII/APPNP | semi-supervised node classification · Cora, CiteSeer, PubMed noisy features | accuracy on noisy nodes | Improved |
| use AirGNN instead of baselines | node classification under Nettack feature attack · Cora, CiteSeer, PubMed adversarial features | accuracy on attacked nodes | Improved |
| use AirGNN instead of baselines | node classification with noisy nodes present · Cora, CiteSeer, PubMed | accuracy on normal nodes | Improved |
| adaptive per-node residual weight (1-beta) for abnormal vs normal nodes (ablation) | AirGNN · Cora noisy features | average residual weight: 0.076 -> 0.002 (normal vs abnormal nodes) | Improved |
Closest papers by meaning
- Residual Connections and Normalization Can Provably Prevent Oversmoothing in GNNs (ICLR 2025 · similarity 0.89)
- Deep Graph Neural Networks via Posteriori-Sampling-based Node-Adaptative Residual Module (NeurIPS 2024 · similarity 0.88)
- Classic GNNs are Strong Baselines: Reassessing GNNs for Node Classification (NeurIPS 2024 · similarity 0.88)
- Can Classic GNNs Be Strong Baselines for Graph-level Tasks? Simple Architectures Meet Excellence (ICML 2025 · similarity 0.88)
- On Vanishing Gradients, Over-Smoothing, and Over-Squashing in GNNs: Bridging Recurrent and Graph Learning (NeurIPS 2025 · similarity 0.88)
- A Simple and Yet Fairly Effective Defense for Graph Neural Networks (AAAI 2024 · similarity 0.87)
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.
