Heterophily-informed Message Passing
Preprint
Abstract
Graph neural networks (GNNs) are known to be vulnerable to oversmoothing due to their implicit homophily assumption. We mitigate this problem with a novel scheme that regulates the aggregation of messages, modulating the type and extent of message passing locally thereby preserving both the low and high-frequency components of information. Our approach relies solely on learnt embeddings, obviating the need for auxiliary labels, thus extending the benefits of heterophily-aware embeddings to broader applications, e.g., generative modelling. Our experiments, conducted across various data sets and GNN architectures, demonstrate performance enhancements and reveal heterophily patterns across standard classification benchmarks. Furthermore, application to molecular generation showcases notable performance improvements on chemoinformatics benchmarks.Keywords
All Related Versions
This publication has 6 references indexed in Scilit:
- Beyond Homophily and Homogeneity Assumption: Relation-Based Frequency Adaptive Graph Neural NetworksIEEE Transactions on Neural Networks and Learning Systems, 2023
- Demystifying Structural Disparity in Graph Neural Networks: Can One Size Fit All?2023
- When Do Graph Neural Networks Help with Node Classification? Investigating the Homophily Principle on Node Distinguishability2023
- Molecular Sets (MOSES): A Benchmarking Platform for Molecular Generation ModelsFrontiers in Pharmacology, 2020
- Automatic Chemical Design Using a Data-Driven Continuous Representation of MoleculesACS Central Science, 2018
- SMILES. 2. Algorithm for generation of unique SMILES notationJournal of Chemical Information and Computer Sciences, 1989