Heterophily-Agnostic Hypergraph Neural Networks with Riemannian Local Exchanger

Researchers propose HealHGNN, a novel Hypergraph Neural Network that addresses limitations in traditional networks when dealing with heterophilic hypergraphs. The system uses Riemannian geometry and adaptive local heat exchangers to enable better long-range dependency modeling with linear complexity.