Dino Duranovic and Ante Javor: From Tables to Graphs: Agentic Graph Modeling скачать в хорошем качестве

Dino Duranovic and Ante Javor: From Tables to Graphs: Agentic Graph Modeling

Relational databases are everywhere, but unlocking their hidden connections often means rethinking them as graphs. In this talk, we’ll explore how AI agents can transform SQL tables into interactive graph models. At the heart of this process is HyGM (Hypothetical Graph Modeling) — an approach that allows large language models to propose, refine, and reshape graph structures. By combining agent workflows with natural language interaction, users can move beyond static schemas and experiment with graph representations that reflect real-world relationships. Once modeled, the data becomes ready for advanced techniques such as GraphRAG, opening new possibilities for search, reasoning, and knowledge discovery.