Quantifying the Knowledge in GNNs for Reliable Distillation into MLPs

TL;DR

This paper introduces a method to improve the distillation of knowledge from GNNs to MLPs by identifying and leveraging reliable knowledge points, leading to significant performance gains across multiple datasets.

Contribution

It proposes a novel reliability measure for GNN knowledge points and a distillation approach that enhances MLP performance by sampling reliable nodes.

Findings

KRD improves MLP accuracy by 12.62%.

KRD outperforms teacher GNNs by 2.16%.

Knowledge point reliability varies across nodes and over time.

Abstract

To bridge the gaps between topology-aware Graph Neural Networks (GNNs) and inference-efficient Multi-Layer Perceptron (MLPs), GLNN proposes to distill knowledge from a well-trained teacher GNN into a student MLP. Despite their great progress, comparatively little work has been done to explore the reliability of different knowledge points (nodes) in GNNs, especially their roles played during distillation. In this paper, we first quantify the knowledge reliability in GNN by measuring the invariance of their information entropy to noise perturbations, from which we observe that different knowledge points (1) show different distillation speeds (temporally); (2) are differentially distributed in the graph (spatially). To achieve reliable distillation, we propose an effective approach, namely Knowledge-inspired Reliable Distillation (KRD), that models the probability of each node being an informative and reliable knowledge point, based on which we sample a set of additional reliable knowledge points as supervision for training student MLPs. Extensive experiments show that KRD improves over the vanilla MLPs by 12.62% and outperforms its corresponding teacher GNNs by 2.16% averaged over 7 datasets and 3 GNN architectures.