Edge-Wise Graph-Instructed Neural Networks

TL;DR

This paper introduces an edge-wise Graph-Instructed (EWGI) layer to enhance Graph-Instructed Neural Networks (GINNs) for multi-task regression, demonstrating improved performance and regularization on complex graph data.

Contribution

It formalizes a novel edge-wise GI layer, addressing limitations of the original GI layer, and shows its advantages through empirical results on different graph types.

Findings

EWGINNs outperform GINNs on Barabasi-Albert graphs.

EWGINNs improve training regularization on Erdos-Renyi graphs.

The edge-wise GI layer enhances multi-task regression on graph data.

Abstract

The problem of multi-task regression over graph nodes has been recently approached through Graph-Instructed Neural Network (GINN), which is a promising architecture belonging to the subset of message-passing graph neural networks. In this work, we discuss the limitations of the Graph-Instructed (GI) layer, and we formalize a novel edge-wise GI (EWGI) layer. We discuss the advantages of the EWGI layer and we provide numerical evidence that EWGINNs perform better than GINNs over some graph-structured input data, like the ones inferred from the Barabasi-Albert graph, and improve the training regularization on graphs with chaotic connectivity, like the ones inferred from the Erdos-Renyi graph.