GraphNNK -- Graph Classification and Interpretability

TL;DR

This paper introduces GraphNNK, a graph classification method that leverages Non-Negative Kernel regression to enhance interpretability and potentially improve generalization in Graph Neural Networks.

Contribution

It proposes integrating NNK with GNNs to provide interpretable predictions based on convex combinations of training examples.

Findings

NNK-based GNNs offer improved interpretability.

The method demonstrates competitive classification performance.

Provides theoretical insights into GNN decision processes.

Abstract

Graph Neural Networks (GNNs) have become a standard approach for learning from graph-structured data. However, their reliance on parametric classifiers (most often linear softmax layers) limits interpretability and sometimes hinders generalization. Recent work on interpolation-based methods, particularly Non-Negative Kernel regression (NNK), has demonstrated that predictions can be expressed as convex combinations of similar training examples in the embedding space, yielding both theoretical results and interpretable explanations.