Skip to main content
Version: 10.3

Explore AutoML SDK components

AutoML SDK is a WorkFusion-built Java framework for creating and customizing machine learning models. It is a powerful tool in addressing classification and information extraction use cases.

Capabilities​

AutoML SDK enables you to accomplish the following tasks:

  • Customize models by configuring AutoML SDK components, such as Annotators, Feature Extractors, and so on.
  • Apply normalization and validation logic and correct OCR errors in terms of post-processing to improve model results.
  • Run model training and publish improved model. 
  • View analytics and statistics on model results.

This toolset delivers a ready-to-use model as it proceeds through the components pipeline.

Components​

AutoML SDK consists of the following components, each responsible for specific tasks in a pipeline:

  • Parser removes HTML tags from documents and creates an AutoML SDK document.
  • Annotators split a text into tokens, add boundary elements and NERs.
  • Feature extractors analyze tokens in documents and create features.
  • ML algorithms aggregate all field documents and feature extractors and provide a model and its statistics.
  • Post-processing applies normalization logic and field grouping.