Learning Robust Heterogeneous Graph Representations via Contrastive-Reconstruction under Sparse Semantics

TL;DR

This paper presents HetCRF, a dual-channel self-supervised framework for heterogeneous graphs that combines contrastive and reconstruction methods, effectively handling sparse semantics and improving node classification performance.

Contribution

HetCRF introduces a two-stage aggregation and positive sample augmentation strategies to unify contrastive and reconstruction learning for heterogeneous graphs.

Findings

HetCRF outperforms state-of-the-art methods on four real-world datasets.

It significantly improves Macro-F1 scores on datasets with missing node features.

The framework effectively addresses semantic sparsity and gradient imbalance issues.

Abstract

In graph self-supervised learning, masked autoencoders (MAE) and contrastive learning (CL) are two prominent paradigms. MAE focuses on reconstructing masked elements, while CL maximizes similarity between augmented graph views. Recent studies highlight their complementarity: MAE excels at local feature capture, and CL at global information extraction. Hybrid frameworks for homogeneous graphs have been proposed, but face challenges in designing shared encoders to meet the semantic requirements of both tasks. In semantically sparse scenarios, CL struggles with view construction, and gradient imbalance between positive and negative samples persists. This paper introduces HetCRF, a novel dual-channel self-supervised learning framework for heterogeneous graphs. HetCRF uses a two-stage aggregation strategy to adapt embedding semantics, making it compatible with both MAE and CL. To address semantic sparsity, it enhances encoder output for view construction instead of relying on raw features, improving efficiency. Two positive sample augmentation strategies are also proposed to balance gradient contributions. Node classification experiments on four real-world heterogeneous graph datasets demonstrate that HetCRF outperforms state-of-the-art baselines. On datasets with missing node features, such as Aminer and Freebase, at a 40% label rate in node classification, HetCRF improves the Macro-F1 score by 2.75% and 2.2% respectively compared to the second-best baseline, validating its effectiveness and superiority.