🤖 AI Summary
Existing heterogeneous graph representation learning methods rely heavily on predefined, coarse-grained meta-paths and struggle to capture fine-grained semantic relationships. To address this, we propose MF2Vec, an unsupervised framework that eliminates manual schema design. Its core innovation lies in a multi-faceted path modeling mechanism: adaptive random walks automatically discover hierarchical semantic paths between nodes, yielding multi-faceted node embeddings; subsequently, homogeneous subgraphs are reconstructed to enable compatibility with shallow embedding techniques (e.g., DeepWalk variants). Evaluated on multiple benchmark datasets, MF2Vec consistently outperforms HGNN-based baselines across node classification, link prediction, and clustering tasks, achieving average improvements of 3.2%–7.8%. The framework offers schema-free flexibility while maintaining strong generalization capability.
📝 Abstract
Recent advancements in graph neural networks (GNNs) and heterogeneous GNNs (HGNNs) have advanced node embeddings and relationship learning for various tasks. However, existing methods often rely on domain-specific predefined meta-paths, which are coarse-grained and focus solely on aspects like node type, limiting their ability to capture complex interactions. We introduce MF2Vec, a model that uses multi-faceted (fine-grained) paths instead of predefined meta-paths. MF2Vec extracts paths via random walks and generates multi-faceted vectors, ignoring predefined schemas. This method learns diverse aspects of nodes and their relationships, constructs a homogeneous network, and creates node embeddings for classification, link prediction, and clustering. Extensive experiments show that MF2Vec outperforms existing methods, offering a more flexible and comprehensive framework for analyzing complex networks. The code is available at https://anonymous.4open.science/r/MF2Vec-6ABC.