Knowledge Probing for Graph Representation Learning

TL;DR

This paper introduces GraphProbe, a framework for analyzing what types of graph properties are encoded in various graph learning methods, using systematic probes and benchmarks across multiple tasks.

Contribution

The paper presents a novel probing framework and comprehensive evaluation benchmark to interpret and compare the knowledge encoding of different graph learning methods.

Findings

GCN and WeightedGCN are versatile across tasks

GraphProbe effectively estimates graph representation capabilities

Different methods encode different levels of graph knowledge

Abstract

Graph learning methods have been extensively applied in diverse application areas. However, what kind of inherent graph properties e.g. graph proximity, graph structural information has been encoded into graph representation learning for downstream tasks is still under-explored. In this paper, we propose a novel graph probing framework (GraphProbe) to investigate and interpret whether the family of graph learning methods has encoded different levels of knowledge in graph representation learning. Based on the intrinsic properties of graphs, we design three probes to systematically investigate the graph representation learning process from different perspectives, respectively the node-wise level, the path-wise level, and the structural level. We construct a thorough evaluation benchmark with nine representative graph learning methods from random walk based approaches, basic graph neural networks and self-supervised graph methods, and probe them on six benchmark datasets for node classification, link prediction and graph classification. The experimental evaluation verify that GraphProbe can estimate the capability of graph representation learning. Remaking results have been concluded: GCN and WeightedGCN methods are relatively versatile methods achieving better results with respect to different tasks.