← Search

Niklas Kemper

2 accepted papers

2025

What Expressivity Theory Misses: Message Passing Complexity for GNNs

NeurIPS 2025spotlight

Expressivity theory, characterizing which graphs a GNN can distinguish, has become the predominant framework for analyzing GNNs, with new models striving for higher expressivity. However, we argue that this focus is misguided: First, higher expressivity is not necessary for most real-world tasks as…

Cited by 0SourceScholar
2024

Expressivity and Generalization: Fragment-Biases for Molecular GNNs

ICML 2024oral

Although recent advances in higher-order Graph Neural Networks (GNNs) improve the theoretical expressiveness and molecular property predictive performance, they often fall short of the empirical performance of models that explicitly use fragment information as inductive bias. However, for these appr…

Cited by 5SourcePDFScholar