Title: Relational.AI Knowledge Graph Management System

Speaker: Martin Bravenboer

Abstract

We present the data model, the Rel language, and reasoning algorithms used in the RelationalAI Knowledge Graph Management System (RAI KGMS). The RAI KGMS uses the relational model for knowledge graphs and implements novel join algorithms to make the relational model effective for reasoning over graphs. The same reasoning and optimization methods are used to express knowledge and end-to-end machine learning problems. We discuss functional and system requirements for a system that manages a knowledge graph.

Slides

Bio

Martin Bravenboer is the VP of Engineering at RelationalAI, where he leads the development of the RelationalAI Knowledge Graph Management System. Before RelationalAI, he was the CTO at LogicBlox. Martin obtained his PhD in Programming Languages at Utrecht University in the Netherlands. During a postdoc, he developed the Doop program analysis system that uses the LogicBlox Datalog engine.