Research map / Graph learning / GNN architectures
Dual Mamba for Node-Specific Representation Learning: Tackling Over-Smoothing with Selective State Space Modeling
Xin He, Yili Wang, Yiwei Dai, Xin Wang
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 GCAMba (global Mamba) (ablation) | DMbaGCN · Pubmed | accuracy (%): 89.67 -> 90.49 | Improved |
| add GCAMba (global Mamba) (ablation) | DMbaGCN · Photo, 16 layers | accuracy (%): 94.28 -> 95.61 | Improved |
| add GCAMba (global Mamba) (ablation) | DMbaGCN · Photo, 2 layers | accuracy (%): 94.42 -> 94.59 | Improved |
| add LSEMba (local state-evolution Mamba) (ablation) | DMbaGCN · CoraFull | accuracy (%): 62.15 -> 72.26 | Improved |
| add LSEMba (local state-evolution Mamba) (ablation) | DMbaGCN · Computers | accuracy (%): 84.18 -> 92.49 | Improved |
| use dual Mamba (LSEMba + GCAMba) instead of local-only MbaGCN (ablation) | DMbaGCN · Pubmed | accuracy (%): 89.32 -> 90.49 | Improved |
| use dual Mamba instead of MbaGCN | DMbaGCN · Photo | accuracy (%): 94.41 -> 95.61 | Improved |
| increase depth from 2 to 32 layers | GCN · Photo | accuracy (%): 93.77 -> 24.03 | Hurt |
| increase depth from 2 to 32 layers | MbaGCN · Photo | accuracy (%): 94.77 -> 91.99 | Hurt |
| increase depth from 2 to 32 layers | DMbaGCN · Photo | accuracy (%): 94.59 -> 95.46 | No clear effect |
| DMbaGCN vs GPRGNN on Physics | DMbaGCN · Physics | accuracy (%): 97.05 -> 97.13 | No clear effect |
| DMbaGCN vs GPRGNN at 2 layers | DMbaGCN · Photo, 2 layers | accuracy (%): 94.77 -> 94.59 | Hurt |
Closest papers by meaning
- GEM: A Dual-Scale Architecture for Graph-Level Hierarchical Representation Learning (NeurIPS 2026 · similarity 0.90)
- DG-Mamba: Robust and Efficient Dynamic Graph Structure Learning with Selective State Space Models (AAAI 2025 · similarity 0.88)
- L2-GCN: Layer-Wise and Learned Efficient Training of Graph Convolutional Networks (CVPR 2020 · similarity 0.87)
- Deep Graph Neural Networks via Posteriori-Sampling-based Node-Adaptative Residual Module (NeurIPS 2024 · similarity 0.87)
- On Vanishing Gradients, Over-Smoothing, and Over-Squashing in GNNs: Bridging Recurrent and Graph Learning (NeurIPS 2025 · similarity 0.86)
- On the Stability of Graph Convolutional Neural Networks: A Probabilistic Perspective (NeurIPS 2025 · 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.
