VeeAlign: Multifaceted Context Representation using Dual Attention for Ontology Alignment

TL;DR

VeeAlign introduces a dual-attention deep learning model for ontology alignment that effectively captures both syntactic and semantic information, demonstrating superior performance across multiple datasets and domains.

Contribution

The paper presents VeeAlign, a novel dual-attention mechanism for ontology alignment that is scalable, domain-agnostic, and leverages contextualized concept representations.

Findings

Outperforms state-of-the-art methods on four diverse datasets

Effectively exploits syntactic and semantic features

Demonstrates scalability and flexibility across domains

Abstract

Ontology Alignment is an important research problem applied to various fields such as data integration, data transfer, data preparation, etc. State-of-the-art (SOTA) Ontology Alignment systems typically use naive domain-dependent approaches with handcrafted rules or domain-specific architectures, making them unscalable and inefficient. In this work, we propose VeeAlign, a Deep Learning based model that uses a novel dual-attention mechanism to compute the contextualized representation of a concept which, in turn, is used to discover alignments. By doing this, not only is our approach able to exploit both syntactic and semantic information encoded in ontologies, it is also, by design, flexible and scalable to different domains with minimal effort. We evaluate our model on four different datasets from different domains and languages, and establish its superiority through these results as well as detailed ablation studies. The code and datasets used are available at https://github.com/Remorax/VeeAlign.