3 papers
cs.LG2026
Do Sheaf Neural Networks Use Holonomy? A Measure--Intervene--Control Study
Ankit Grover, Rémi Bourgerie, Rémi Bourgerie
Geometric architectures are often motivated by internal mechanisms, but accuracy alone does not show whether predictions use them. In Sheaf Neural Networks (SNNs), edge transports…
cs.LG2026
Deep Neural Sheaf Diffusion
Rémi Bourgerie, Šarūnas Girdzijauskas, Viktoria Fodor
Deep Graph Neural Networks (GNNs) are essential for capturing complex dependencies in graph-structured data. However, scaling GNNs to depth remains challenging, as stacking layers…
cs.LG2026
Is One Token All It Takes? Graph Pooling Tokens for LLM-based GraphQA
Ankit Grover, Lodovico Giaretta, Rémi Bourgerie +1
The integration of Graph Neural Networks (GNNs) with Large Language Models (LLMs) has emerged as a promising paradigm for Graph Question Answering (GraphQA). However, effective met…