Graph Topology Information Enhanced Heterogeneous Graph Representation Learning

2026-04-07

AI Summary

Real-world heterogeneous graphs are often noisy and structurally suboptimal, which hinders the performance of graph representation learning. Existing graph structure learning (GSL) approaches are predominantly designed for homogeneous graphs and struggle to generalize to heterogeneous settings. To address this gap, this work proposes ToGRL, a novel framework that effectively extends GSL to heterogeneous graphs for the first time. ToGRL employs a two-stage decoupled strategy: it first leverages a new GSL module to extract task-relevant topological information and constructs an optimized graph structure with smoother signals; it then performs representation learning on this refined graph and integrates prompt tuning to enhance adaptability to downstream tasks. By jointly exploiting topological embeddings and prompt tuning, ToGRL significantly outperforms state-of-the-art methods across five real-world datasets while reducing memory overhead and improving both model expressiveness and generalization capability.

Abstract

Real-world heterogeneous graphs are inherently noisy and usually not in the optimal graph structures for downstream tasks, which often adversely affects the performance of GRL models in downstream tasks. Although Graph Structure Learning (GSL) methods have been proposed to learn graph structures and downstream tasks simultaneously, existing methods are predominantly designed for homogeneous graphs, while GSL for heterogeneous graphs remains largely unexplored. Two challenges arise in this context. Firstly, the quality of the input graph structure has a more profound impact on GNN-based heterogeneous GRL models compared to their homogeneous counterparts. Secondly, most existing homogenous GRL models encounter memory consumption issues when applied directly to heterogeneous graphs. In this paper, we propose a novel Graph Topology learning Enhanced Heterogeneous Graph Representation Learning framework (ToGRL).ToGRL learns high-quality graph structures and representations for downstream tasks by incorporating task-relevant latent topology information. Specifically, a novel GSL module is first proposed to extract downstream task-related topology information from a raw graph structure and project it into topology embeddings. These embeddings are utilized to construct a new graph with smooth graph signals. This two-stage approach to GSL separates the optimization of the adjacency matrix from node representation learning to reduce memory consumption. Following this, a representation learning module takes the new graph as input to learn embeddings for downstream tasks. ToGRL also leverages prompt tuning to better utilize the knowledge embedded in learned representations, thus enhancing adaptability to downstream tasks. Extensive experiments on five real-world datasets show that our ToGRL outperforms state-of-the-art methods by a large margin.

Problem

• heterogeneous graph
• graph structure learning
• graph representation learning
• noisy graph
• memory consumption

Innovation

• Heterogeneous Graph
• Graph Structure Learning
• Topology Embedding
• Prompt Tuning
• Graph Representation Learning