AutoML Natural Language Entity Extraction offers support for document-based entity extraction, which enables you to identify entities in a format that recreates the basic layout of the training item

## Feature AutoML Natural Language Entity Extraction offers support for document-based entity extraction, which enables you to identify entities in a format that recreates the basic layout of the training item. When you add annotations, AutoML Natural Language Entity Extraction captures the annotation's position on the page as a factor considered during training. See [Managing datasets](https://cloud.google.com/natural-language/automl/entity-analysis/docs/datasets#label-items) for more information.