Leiden-Fusion Partitioning Method for Effective Distributed Training of Graph Embeddings

TL;DR

Leiden-Fusion is a novel graph partitioning method that creates densely connected subgraphs for distributed GNN training, reducing communication overhead and maintaining embedding quality.

Contribution

We introduce Leiden-Fusion, a partitioning algorithm that ensures connected subgraphs for efficient distributed graph embedding training, addressing key limitations of existing methods.

Findings

Reduces communication in distributed GNN training

Maintains high embedding quality on benchmark datasets

Ensures partitions are densely connected without isolated nodes

Abstract

In the area of large-scale training of graph embeddings, effective training frameworks and partitioning methods are critical for handling large networks. However, they face two major challenges: 1) existing synchronized distributed frameworks require continuous communication to access information from other machines, and 2) the inability of current partitioning methods to ensure that subgraphs remain connected components without isolated nodes, which is essential for effective training of GNNs since training relies on information aggregation from neighboring nodes. To address these issues, we introduce a novel partitioning method, named Leiden-Fusion, designed for large-scale training of graphs with minimal communication. Our method extends the Leiden community detection algorithm with a greedy algorithm that merges the smallest communities with highly connected neighboring communities. Our method guarantees that, for an initially connected graph, each partition is a densely connected subgraph with no isolated nodes. After obtaining the partitions, we train a GNN for each partition independently, and finally integrate all embeddings for node classification tasks, which significantly reduces the need for network communication and enhances the efficiency of distributed graph training. We demonstrate the effectiveness of our method through extensive evaluations on several benchmark datasets, achieving high efficiency while preserving the quality of the graph embeddings for node classification tasks.