Abstract:Recent years have witnessed great success in handling graph-related tasks with graph neural networks (GNNs). However, most existing GNNs are based on message passing to perform feature aggregation and transformation, where the structural information is explicitly involved in the forward propagation by coupling with node features through graph convolution at each layer. As a result, subtle feature noise or structure perturbation may cause severe error propagation, resulting in extremely poor robustness. In this article, we rethink the roles played by graph structural information in graph data training and identify that message passing is not the only path to modeling structural information. Inspired by this, we propose a simple but effective graph structure self-contrasting (GSSC) framework that learns graph structural information without message passing. The proposed framework is based purely on multilayer perceptrons (MLPs), where the structural information is only implicitly incorporated as prior knowledge to guide the computation of supervision signals, substituting the explicit message propagation as in GNNs. Specifically, it first applies structural sparsification (STR-Sparse) to remove potentially uninformative or noisy edges in the neighborhood, and then performs structural self-contrasting (STR-Contrast) in the sparsified neighborhood to learn robust node representations. Finally, STR-Sparse and self-contrasting are formulated as a bilevel optimization problem and solved in a unified framework. Extensive experiments have qualitatively and quantitatively demonstrated that the GSSC framework can produce truly encouraging performance with better generalization and robustness than other leading competitors. Codes are publicly available at: https://github.com/LirongWu/GSSC.

Graph Neural Network Based Multi-Instance Learning with Graph Structure Learning

LoSS: Local Structural Separation Hypergraph Convolutional Neural Network

Hierarchical Multi-View Graph Pooling with Structure Learning

Semi-Supervised Graph Embedding Via Multi-instance Kernel Learning

Multi-task Learning over Graph Structures.

Multi-View Subgraph Neural Networks: Self-Supervised Learning with Scarce Labeled Data

Supervised Contrastive Learning with Structure Inference for Graph Classification

Local Structure-aware Graph Contrastive Representation Learning

Semi-Supervised and Self-Supervised Classification with Multi-View Graph Neural Networks

Learning to Model Graph Structural Information on MLPs via Graph Structure Self-Contrasting

Generative Essential Graph Convolutional Network for Multi-View Semi-Supervised Classification

Improving Subgraph Representation Learning via Multi-View Augmentation

Self-restrained contrastive enhanced network for graph structure learning

Towards Unsupervised Deep Graph Structure Learning

GraphSAGE++: Weighted Multi-scale GNN for Graph Representation Learning

Multi-Modal Multi-Instance Multi-Label Learning with Graph Convolutional Network

Line graph contrastive learning for node classification

MeGraph: Graph Representation Learning on Connected Multi-scale Graphs

Neural Networks for Multi-Instance Learning

Self-Attention Empowered Graph Convolutional Network for Structure Learning and Node Embedding

MDS-GNN: A Mutual Dual-Stream Graph Neural Network on Graphs with Incomplete Features and Structure