Automatic Data Classification – AI that continuously learns from human input without supervised learning

In this tutorial, we walk through how DvSum’s AI-powered data classification engine continuously learns from human input to instantly improve the accuracy of its data classification, without the need for supervised learning.

Context:
An important requirement in a hybrid data landscape is linking physical assets like datasets and fields that are created in separate applications to a consistent unified and business context enriched definition. Why? Because it allows for a more effective search of data by technical and non-technical users, and it allows for applying standards to data to monitor and improve its quality.

Business Glossary and Data Taxonomies can be used to manage those definitions. However, the challenge is the ownership and the manual effort required to review and link the physical data assets to these definitions.

Neither of the System and database administrators of application, analytics teams, and business users see as their job. Some organizations that have a dedicated data management function and/or data governance program will have data stewards, one of their jobs to link and tag the physical datasets with glossary terms. Additionally, manual curation is impractical and unsustainable as the data landscape and volume continue to grow.

With a powerful AI-based classification engine, DvSum’s ML Data Catalog automatically catalogs, classifies, and curates data from your numerous data siloes and makes it available as a unified, actionable Data Catalog. DvSum’s AI-powered data classification engine continuously learns from human input to instantly improve the accuracy of its data classification, without the need for supervised learning.