Machine-readable regulations and large language models

The Norwegian Maritime Authority has since 2020 established machine-readable regulatory requirements, based on open standards for semantic knowledge graphs. Machine-readable regulations are at core in the implementation of a new supervision system. The implementation of machine-readable regulations relies on Natural Language Processing and RDF transformations. This talk will walk you through the thought process, pipeline and results. And why we chose SHACL for modelling regulatory requirements. We will also cover a small experiment of using Large Language Models (LLM) on information extraction from the knowledge graph.

Speakers:
Veronika Heimsbakk, Knowledge Graph Lead at Capgemini
Sebastian Amundsen, Knowledge Graph Nerd at Capgemini