3 papers
cs.LG2026
Geometric Structures on Graphs: a Holonomy-Based Discretization of Curvature
Hao Li, Yuhan Peng, Junwen Dong
We propose a holonomy-based framework for discretizing curvature on graphs equipped with local symmetric positive-definite metrics. Each vertex carries a fibre metric \(g_i\), and…
cs.LG2026
Demystifying Oversmoothing in Sheaf Neural Networks: An Index-Theoretic Criterion
Junwen Dong, Yuhan Peng, Hao Li +2
To combat oversmoothing in Graph Convolutional Networks, Sheaf Neural Networks (SNNs) were proposed as a generalization by equipping the graph with a sheaf structure and replacing…
cs.LG2026
Sheaf Neural Networks on SPD Manifolds: Second-Order Geometric Representation Learning
Yuhan Peng, Junwen Dong, Yuzhi Zeng +6
Graph neural networks face two fundamental challenges rooted in the linear structure of Euclidean vector spaces: (1) Current architectures represent geometry through vectors (direc…