Ordered Topological Deep Learning: A Network Modeling Case Study

Bernárdez. G. , Ferriol-Galmés, M., Güemes-Palau, C., Papillon, M., Barlet-Ros, P., Cabellos-Aparicio, A., Miolane, N.

Computer networks underpin modern digital infrastructure, but traditional modeling methods often fall short due to high computational cost or restrictive assumptions. Deep learning–based models such as RouteNet have redefined network modeling, offering a powerful balance between accuracy and efficiency. In this work, we reveal RouteNet’s hidden foundation in Topological Deep Learning and introduce OrdGCCN, a novel ordered topological framework that demonstrates the first successful real-world application of TDL principles.