A Flexible Generative Framework for Graph-based Semi-supervised Learning

TL;DR

This paper introduces a flexible generative framework for semi-supervised learning on graph-structured data, leveraging joint distributions of features, labels, and graph structure, and demonstrates superior performance over existing methods.

Contribution

It proposes a novel generative approach that models the joint distribution of features, labels, and graph structure, improving semi-supervised learning performance.

Findings

Outperforms state-of-the-art models on benchmark datasets

Utilizes scalable variational inference for label prediction

Effectively leverages relational information in graphs

Abstract

We consider a family of problems that are concerned about making predictions for the majority of unlabeled, graph-structured data samples based on a small proportion of labeled samples. Relational information among the data samples, often encoded in the graph/network structure, is shown to be helpful for these semi-supervised learning tasks. However, conventional graph-based regularization methods and recent graph neural networks do not fully leverage the interrelations between the features, the graph, and the labels. In this work, we propose a flexible generative framework for graph-based semi-supervised learning, which approaches the joint distribution of the node features, labels, and the graph structure. Borrowing insights from random graph models in network science literature, this joint distribution can be instantiated using various distribution families. For the inference of missing labels, we exploit recent advances of scalable variational inference techniques to approximate the Bayesian posterior. We conduct thorough experiments on benchmark datasets for graph-based semi-supervised learning. Results show that the proposed methods outperform the state-of-the-art models in most settings.