PulseAugur
LIVE 13:10:33
research · [2 sources] ·
0
research

Topological Neural Tangent Kernel enhances graph neural networks with higher-order structure

Researchers have introduced the Topological Neural Tangent Kernel (TopoNTK), a novel kernel designed for simplicial message passing that extends beyond pairwise relationships. Unlike traditional graph kernels, TopoNTK can capture higher-order interactions within simplicial complexes, making it sensitive to topological structures invisible to graph-based methods. This approach offers a more interpretable learning geometry by decomposing edge signals and analyzing how different components are learned based on spectral properties. AI

Summary written by gemini-2.5-flash-lite from 2 sources. How we write summaries →

IMPACT Introduces a new kernel for graph neural networks that captures higher-order interactions, potentially improving relational learning interpretability and effectiveness.

RANK_REASON This is a research paper introducing a new kernel for graph neural networks.

Read on arXiv stat.ML →

COVERAGE [2]

  1. arXiv stat.ML TIER_1 Italiano(IT) · Sanjukta Krishnagopal ·

    Topological Neural Tangent Kernel

    arXiv:2605.01110v1 Announce Type: cross Abstract: Graph neural tangent kernels give a principled infinite-width theory for graph neural networks, but inherit a basic limitation of graph models: they see only pairwise structure. Many relational systems contain higher-order interac…

  2. arXiv stat.ML TIER_1 Italiano(IT) · Sanjukta Krishnagopal ·

    Topological Neural Tangent Kernel

    Graph neural tangent kernels give a principled infinite-width theory for graph neural networks, but inherit a basic limitation of graph models: they see only pairwise structure. Many relational systems contain higher-order interactions that are more naturally represented by simpl…