Nonlinear Correct and Smooth for Semi-Supervised Learning

TL;DR

This paper introduces NLCS, a nonlinear method for semi-supervised learning on graphs that enhances label propagation by incorporating higher-order relationships and non-linearity, leading to significant performance improvements.

Contribution

The paper proposes NLCS, a novel nonlinear residual propagation method that effectively leverages labels and features in higher-order graphs for improved semi-supervised learning.

Findings

Achieves 13.71% average improvement over base predictions.

Outperforms state-of-the-art post-processing methods by 2.16%.

Effectively utilizes higher-order graph relationships.

Abstract

Graph-based semi-supervised learning (GSSL) has been used successfully in various applications. Existing methods leverage the graph structure and labeled samples for classification. Label Propagation (LP) and Graph Neural Networks (GNNs) both iteratively pass messages on graphs, where LP propagates node labels through edges and GNN aggregates node features from the neighborhood. Recently, combining LP and GNN has led to improved performance. However, utilizing labels and features jointly in higher-order graphs has not been explored. Therefore, we propose Nonlinear Correct and Smooth (NLCS), which improves the existing post-processing approach by incorporating non-linearity and higher-order representation into the residual propagation to handle intricate node relationships effectively. Systematic evaluations show that our method achieves remarkable average improvements of 13.71% over base prediction and 2.16% over the state-of-the-art post-processing method on six commonly used datasets. Comparisons and analyses show our method effectively utilizes labels and features jointly in higher-order graphs to resolve challenging graph relationships.