Abstract:Graph self-supervised representation learning has gained considerable attention and demonstrated remarkable efficacy in extracting meaningful representations from graphs, particularly in the absence of labeled data. Two representative methods in this domain are graph auto-encoding and graph contrastive learning. However, the former methods primarily focus on global structures, potentially overlooking some fine-grained information during reconstruction. The latter methods emphasize node similarity across correlated views in the embedding space, potentially neglecting the inherent global graph information in the original input space. Moreover, handling incomplete graphs in real-world scenarios, where original features are unavailable for certain nodes, poses challenges for both types of methods. To alleviate these limitations, we integrate masked graph auto-encoding and prototype-aware graph contrastive learning into a unified model to learn node representations in graphs. In our method, we begin by masking a portion of node features and utilize a specific decoding strategy to reconstruct the masked information. This process facilitates the recovery of graphs from a global or macro level and enables handling incomplete graphs easily. Moreover, we treat the masked graph and the original one as a pair of contrasting views, enforcing the alignment and uniformity between their corresponding node representations at a local or micro level. Lastly, to capture cluster structures from a meso level and learn more discriminative representations, we introduce a prototype-aware clustering consistency loss that is jointly optimized with the above two complementary objectives. Extensive experiments conducted on several datasets demonstrate that the proposed method achieves significantly better or competitive performance on downstream tasks, especially for graph clustering, compared with the state-of-the-art methods, showcasing its superiority in enhancing graph representation learning.

Cross-View Masked Model for Self-Supervised Graph Representation Learning

SGLCMR: Self-supervised Graph Learning of Generalized Representations for Cross-Market Recommendation.

Efficient self-supervised heterogeneous graph representation learning with reconstruction

GraphMAE2: A Decoding-Enhanced Masked Self-Supervised Graph Learner

Self-supervised Graph Representations with Generative Adversarial Learning

Masked Graph Modeling with Multi- View Contrast

ProtoMGAE: Prototype-aware Masked Graph Auto-Encoder for Graph Representation Learning

Multi-view Robust Graph Representation Learning for Graph Classification

GMCL: Graph Mask Contrastive Learning for Self-Supervised Graph Representation Learning

What's Behind the Mask: Understanding Masked Graph Modeling for Graph Autoencoders

Graph Contrastive Learning with Cross-view Reconstruction

RARE: Robust Masked Graph Autoencoder

Heterogeneous Graph Masked Autoencoders

UGMAE: A Unified Framework for Graph Masked Autoencoders

SeeGera: Self-supervised Semi-implicit Graph Variational Auto-encoders with Masking

Contrastive Graph Representation Learning with Adversarial Cross-view Reconstruction and Information Bottleneck

Multiview Subgraph Neural Networks: Self-Supervised Learning With Scarce Labeled Data

Unsupervised Multi-View Graph Representation Learning with Dual Weight-Net

Self-supervised Graph Learning for Recommendation

Disentangled Generative Graph Representation Learning

GraphMAE: Self-Supervised Masked Graph Autoencoders