Abstract:Recent years have witnessed the rapid development of heterogeneous graph neural networks (HGNNs) in information retrieval (IR) applications. Many existing HGNNs design a variety of tailor-made graph convolutions to capture structural and semantic information in heterogeneous graphs. However, existing HGNNs usually represent each node as a single vector in the multi-layer graph convolution calculation, which makes the high-level graph convolution layer fail to distinguish information from different relations and different orders, resulting in the information loss in the message passing. %insufficient mining of information. To this end, we propose a novel heterogeneous graph neural network with sequential node representation, namely Seq-HGNN. To avoid the information loss caused by the single vector node representation, we first design a sequential node representation learning mechanism to represent each node as a sequence of meta-path representations during the node message passing. Then we propose a heterogeneous representation fusion module, empowering Seq-HGNN to identify important meta-paths and aggregate their representations into a compact one. We conduct extensive experiments on four widely used datasets from Heterogeneous Graph Benchmark (HGB) and Open Graph Benchmark (OGB). Experimental results show that our proposed method outperforms state-of-the-art baselines in both accuracy and efficiency. The source code is available at <a class="link-external link-https" href="https://github.com/nobrowning/SEQ_HGNN" rel="external noopener nofollow">this https URL</a>.

What problem does this paper attempt to address?

The paper primarily addresses the limitations of Heterogeneous Graph Neural Networks (HGNNs) in Information Retrieval (IR) applications by proposing a new solution. ### Research Background and Problem Existing HGNNs typically represent each node as a single vector in multi-layer graph convolution calculations, which makes it difficult for higher-layer graph convolutional layers to distinguish information from different relationships and orders, leading to information loss. For example, in a heterogeneous graph containing different types of nodes (such as movies, directors, keywords, and actors), existing methods may mix information from these different types of nodes and fail to distinguish them. ### Solution To address the above issues, the authors propose Seq-HGNN, a novel heterogeneous graph neural network that utilizes sequential node representations to avoid information loss and achieves this through the following two key steps: 1. **Sequential Node Representation Learning Mechanism**: During message passing, Seq-HGNN represents each node as a series of meta-path representations instead of a single vector. This allows it to record messages passed along multiple meta-paths, maintaining the integrity of structural information. 2. **Heterogeneous Representation Fusion Module**: To further enhance model performance, Seq-HGNN designs a heterogeneous representation fusion module to identify important meta-paths and aggregate their representations into a compact form for downstream tasks. ### Experimental Results Experiments on four widely used datasets (including datasets from the Heterogeneous Graph Benchmark (HGB) and Open Graph Benchmark (OGB)) validate that Seq-HGNN outperforms existing methods in both accuracy and efficiency. Particularly on the DBLP and ACM datasets, Seq-HGNN shows significant improvements in macro-f1 and micro-f1 metrics. ### Conclusion In summary, the main contribution of this paper is the proposal of a new heterogeneous graph neural network—Seq-HGNN, which overcomes the information loss problem of existing methods by introducing sequential node representations and demonstrates superior performance in the field of information retrieval through experimental validation.

Seq-HGNN: Learning Sequential Node Representation on Heterogeneous Graph

Heterogeneous Graph Neural Network With Multi-View Representation Learning

Simple and Efficient Heterogeneous Graph Neural Network

HGBER: Heterogeneous Graph Neural Network With Bidirectional Encoding Representation

MHGNN: Multi-view fusion based Heterogeneous Graph Neural Network

OSGNN: Original graph and Subgraph aggregated Graph Neural Network

Semantic-guided graph neural network for heterogeneous graph embedding

HybridGNN: Learning Hybrid Representation in Multiplex Heterogeneous Networks

Heterogeneous Deep Graph Infomax

SR-HGN: Semantic- and Relation-Aware Heterogeneous Graph Neural Network

DiffMG: Differentiable Meta Graph Search for Heterogeneous Graph Neural Networks

Graph neural architecture search with heterogeneous message-passing mechanisms

SHGNN: Structure-Aware Heterogeneous Graph Neural Network

PHGNN: Position-aware Graph Neural Network for Heterogeneous Graph Embedding.

Topic-aware Heterogeneous Graph Neural Network for Link Prediction

Heterogeneous Graph Representation Learning with Relation Awareness

RHGNN: Imposing Relational Inductive Bias for Heterogeneous Graph Neural Network

HGAMLP: Heterogeneous Graph Attention MLP with De-Redundancy Mechanism

Heterogeneous Graph Propagation Network

Heterogeneous graph neural networks analysis: a survey of techniques, evaluations and applications

Heterogeneous Graph Neural Networks using Self-supervised Reciprocally Contrastive Learning