Compositional PAC-Bayes: Generalization of GNNs with Persistence and Beyond

We study generalization guarantees for graph neural networks with persistence-based representations through a compositional PAC-Bayes perspective.

Recommended citation: Kirill Brilliantov, Amauri H. Souza, and Vikas Garg. "Compositional PAC-Bayes: Generalization of GNNs with Persistence and Beyond." NeurIPS, 2024.
Download Paper