Serving up-to-date Dynamic Knowledge Graph Embeddings

This project addresses the challenge of providing up-to-date embeddings for dynamic knowledge graphs (KGs) and ontologies. The goal is to develop a framework capable of periodically collecting new KG versions, computing embeddings and making them publicly available for further research.

Providing publicly available up-to-date embeddings will:

• Democratize researchers’ access
• Empower researchers to faster experimentation
• Improve performance across different tasks over KGs
• Reduce the carbon footprint by reusing pre-trained embeddings

The project will be funded by an 2024 Open Science Grant awarded to Dr. Rita Torres de Sousa.