Skip to main content
Version: 10.3.2

Introduction to AutoML

AutoML is a proprietory WorkFusion technology that supports machine learning (ML) tasks within the Work.AI platform. The article describes how to work with the AutoML technology and gives an overview of typical uses cases.

tip

For additional information about machine learning and related concepts, read Learn ML basics.

Ways to work with AutoML​

  • AutoML SDK

    The WorkFusion Java framework is intended for expert developer users to create and customize ML models by coding. For details, read the topics in the Explore AutoML SDK components and Get started with AutoML SDK sections.

  • AutoML UI

    The user interface in Control Tower is designed for no-code manipulations with AutoML SDK and LayoutLM models and pipelines:

    • Training ML models

    • Creating and labeling datasets

    • Reviewing model experiment data

    For details, read the topics in the Manage AutoML models in Control Tower section.

  • ML Ops Platform

    The optional Work.AI add-in, built on top of Kubernetes, features the functionality to develop and train ML models. ML Ops Platform is language- and framework-agnostic and gives expert developers access to a larger pool of computational resources. It allows reducing model training time and working with more powerful models, such as based on the LayoutLM deep learning pipeline. For details, read Work with ML Ops Platform.

Typical use cases​

AutoML is used for, but not limited to, the following use cases:

Information extraction​

Information extraction (IE) is when data defined by business logic is taken out (extracted) from unstructured or semi-structured documents and processed according to business rules. In terms of information extraction, each data point is referred to as a field.

For example, for invoices, information extraction might cover such data points as invoice number, supplier name, and quantity of products. It means the three fields to be extracted from documents are invoice number, supplier name, and quantity.

For more information, read the Information extraction topic.

Classification​

Classification means arranging data into groups or categories according to established criteria. Both the binary and multi-class types are supported. It is applied when it is necessary to define the class for an item (document). Classes stand for different document types.

For example, one business flow can include invoices, purchase orders, and claims, but each document type must be handled differently. This means that, before applying automation, you need to classify these documents.

For more information, read the Classification topic.

More automation opportunities​

The following video covers typical use cases automated in insurance, healthcare, financial services, and other sectors. It also describes what problems are addressed by WorkFusion and how they are solved using its Work.AI platform.