From Data to Metadata for Machine Learning Platforms

In this special guest feature, Jörg Schad, Head of Machine Learning at ArangoDB, discusses discuss the need for Machine Learning Metadata, solutions for storing and analyzing Metadata as well as the benefits for the different stakeholders. We all know good training data is crucial for data scientists to build quality machine learning models. But when productionizing Machine Learning, Metadata from the Machine Learning Platform is equally important providing for example:
provenance of model allowing for reproducible builds and audit trails;
context to comply with GDPR, CCPA requirements; and identifying data shift in your production data.