The Danish Business Authority uses machine learning and graph networks to improve its services and oversight of businesses. Its ML Lab team of 9 people enriches a knowledge graph containing business data with machine learning insights. This allows the authority to automate intelligent controls, such as identifying fictional assets or unpaid taxes. It aims to handle complex relationships and automate decisions using a data platform that connects registry data to metadata, observations from machine learning models, and a knowledge graph. Challenges include governance of ML data, explainability of automated decisions, and business transformation. The authority is working to apply these techniques while maintaining transparency, fairness, and accountability.
Related topics: