KERMIT -- A Transformer-Based Approach for Knowledge Graph Matching

TL;DR

KERMIT introduces a scalable transformer-based method for knowledge graph matching that combines bi-encoder candidate generation with cross-encoder refinement, achieving competitive results across multiple datasets.

Contribution

The paper presents a novel two-step transformer-based approach that improves scalability and accuracy in knowledge graph matching compared to existing methods.

Findings

Scalable matching with quadratic complexity reduction.

Competitive accuracy on multiple datasets.

Effective combination of bi-encoder and cross-encoder models.

Abstract

One of the strongest signals for automated matching of knowledge graphs and ontologies are textual concept descriptions. With the rise of transformer-based language models, text comparison based on meaning (rather than lexical features) is available to researchers. However, performing pairwise comparisons of all textual descriptions of concepts in two knowledge graphs is expensive and scales quadratically (or even worse if concepts have more than one description). To overcome this problem, we follow a two-step approach: we first generate matching candidates using a pre-trained sentence transformer (so called bi-encoder). In a second step, we use fine-tuned transformer cross-encoders to generate the best candidates. We evaluate our approach on multiple datasets and show that it is feasible and produces competitive results.