Supervision Tasks

Supervision Tasks

What is a Nested table?

Overview A nested table is a data structure that represents a table within another table. It is embedded as a row into another table, creating a hierarchical or a “nested” structure. Nested tables allow you to extract data from tables with complicated relationships.

Manual Rotation

Overview Sometimes, your documents may go to Supervision with an incorrect orientation. An improperly rotated document may be impossible to read or extract data from. The Manual Rotation feature allows you to correct the orientation of a page within the Supervision console.

Reprocessing

Document Classification is used to categorize and combine pages that were not classified by the machine. Learn more in Document Classification. The document-organization process allows users to recover mismatched documents by requesting reclassification.

Knowledge Store

If you have Custom Supervision tasks in your flows, and you would like your keyers to be able to select from a validated list of items in a particular Custom Supervision task, you can store those items in the Knowledge Store in the Hyperscience application.

What is Supervision?

A submission will travel through multiple steps during the data extraction process. At each step, if the machine's confidence falls below the threshold specified in the flow settings or in Administration > System Settings, the system will trigger a Supervision task.

Document Classification

The Document Classification Supervision task is the first step in Extraction Supervision. It is used to categorize and combine pages that were not classified by the machine. To enable Document Classification, select the Manual Classification Supervision task.

Document Classification Model Validation Tasks

During Classification Model training, the model may classify a page with high confidence but be incorrect. This kind of error is called a high-confidence error. In order to verify a high-confidence error was made and correct the classification, use the validation tasks.

Field Identification

Field Identification (Field ID) tasks are specific to Semi-structured documents. These tasks are created when the system is unsure if it has correctly located a field, or if a field is specified for manual identification in the Layout Editor.

Best Practices for Field ID Supervision and QA

The following best practices for Field Identification Supervision and Quality Assurance will ensure the highest levels of accuracy and automation for your Semi-structured documents. They will also maximize the quality of the training data that is collected.

Table Identification

Overview A Table is a data structure used to organize and present information in rows and columns. It is used to present values in a readable format. A table consists of the following elements: Rows, Columns, Cells. Hyperscience provides tools to facilitate table identification during data extraction.

Relationship Between Field ID and Table ID

Overview Documents can contain normal fields, table fields, or both. The Supervision tasks associated with a document are dependent on the types of fields it contains. In each of the following scenarios, the field type(s) in a document are predetermined and can dictate the supervision tasks executed.

Transcription

Your flow can have the following types of Transcription tasks: Field Transcription, Table Transcription, and Flexible Extraction. A Field or Table Transcription task can be generated in the following ways: The machine's transcription is believed to fall short of accuracy.

Transcription Supervision Consensus

To achieve Transcription Supervision Consensus for a given field, two transcription entries must have matching normalized transcriptions. At least one manual transcription of the given field will be required, regardless of the machine’s confidence level.

Best Practices for Table Transcription

Table Transcription During Table Transcription, keyers can sometimes have difficulties finding cells that are pending transcription or finding transcribed cells that need to be reviewed. To save time when navigating to different cells, we recommend enabling cell highlighting.

Configuring Custom Rejection Reasons

Overview With the “Reject” button, you can stop processing documents that are identified as not in good order. Note that this button will only appear during Transcription Supervision for Structured documents. To reject a document, go through the guided process provided.

Custom Supervision

Overview With the information available in our Flows SDK and the help of your Hyperscience representative, you can include Custom Supervision in your flows. Custom Supervision allows you to tailor a Supervision task’s interface to a specific need.

Text Classification

With Hyperscience’s Text Classification feature, you can train a model to classify freeform text in documents, emails, and more. This feature allows you to analyze and organize unstructured text by user intent, sentiment, topic, or any custom label.

Data Lookup and Validation with Flow Blocks

Overview To minimize Supervision and increase automation, you can use data lookups and validations. Each flow can be configured to pull and validate data, using Custom Code Blocks, Database Blocks, and API Blocks. To learn more about these block configurations, refer to the relevant sections in the documentation.