Knowledge Graph Enhanced Schema Matching Network for Heterogeneous Data Integration

Abstract

The semantic heterogeneity in various schemas brings some obstacles while integrating and establishing the correspondences between source schema and target schema, hence human interventions and domain knowledge are required to tackle some complex mapping tasks for heterogeneous data integration. In this presentation, a knowledge graph enhanced schema matching network for heterogeneous data integration is introduced to reduce the human intervention in schema matching without losing accuracy for complex schema mapping. Specifically, the knowledge representation learning (i.e., OWL2Vec, KGs embedding, etc.) is employed to enhance the traditional schema matching network by knowledge injection. With the help of knowledge injection, the common knowledge and hidden semantics among the database elements could be employed to tackle the semantic heterogeneity of the schema mapping, e.g., abbreviations, acronyms, synonyms, etc. which is suitable for complex mapping.