A Collection of Latest Interesting Papers on Graph Neural Networks

NeurIPS 2025

1. A Signed Graph Approach to Understanding and Mitigating Oversmoothing

2. Relieving the Over-Aggregating Effect in Graph Transformers

3. Effects of Dropout on Performance in Long-range Graph Learning Tasks

4. Deeper with Riemannian Geometry: Overcoming Oversmoothing and Oversquashing for Graph Foundation Models

5. TopER: Topological Embeddings in Graph Representation Learning

6. Graph Diffusion that can Insert and Delete

7. Generalizable Insights for Graph Transformers in Theory and Practice

8. A Closer Look at Graph Transformers: Cross-Aggregation and Beyond

9. Bridging Theory and Practice in Link Representation with Graph Neural Networks

10. Over-squashing in Spatiotemporal Graph Neural Networks

11. Robust Explanations of Graph Neural Networks via Graph Curvatures

12. HubGT: Fast Graph Transformer with Decoupled Hierarchy Labeling

13. MultiNet: Adaptive Multi-Viewed Subgraph Convolutional Networks for Graph Classification

14. OpenGU: A Comprehensive Benchmark for Graph Unlearning

15. Depth-Width Tradeoffs for Transformers on Graph Tasks

16. Spectral Graph Coarsening Using Inner Product Preservation and the Grassmann Manifold

17. Rethinking Tokenized Graph Transformers for Node Classification

18. Topology-aware Graph Diffusion Model with Persistent Homology

19. Enhancing Graph Classification Robustness with Singular Pooling

20. Memorization in Graph Neural Networks

21. Pseudo-Riemannian Graph Transformer