Aurora for Instant Learning
Feature of Aurora AI that allows it to learn instantly from documents annotated by the users (human feedback) and use gathered knowledge to predict values on documents in the future.
Instant learning (legacy)
A basic learning feature that enables Rossum to learn quickly based on customers’ previous annotations. It recognizes the document’s confirmed layout and then applies predictions based on this information to other documents with the same format.
Extensions
The Rossum Extensions Environment effectively automates workflows and can provide new capabilities using pre-built or custom solutions.
For example, you can route documents to different queues based on your requirements, match extracted data with a database, or quickly identify duplicates.
Automation blocker
Admin and Manager level users can see the automation blockers on the validation screen. They appear when Rossum’s engine finds fields that block automation process.
Tab
A part of the main user interface. In Rossum, multiple tabs are available that relate to the status of a document for a given queue. That includes "To review", "Postponed", "Exported", "Rejected", "Confirmed", and "Deleted" tabs.
Furthermore, the Emails tab allows you to track all documents imported to Rossum’s email inbox.
Usage reporting dashboard
Available from the Statistics tab, it allows users to explore document-related information for all queues and workspaces.
Schema Editor
A section of Rossum where you can modify the extraction schema. You can specify which data fields should be captured in a specific queue. Custom fields and sections can also be added to allow the user to extract all required data.
Bounding box
A rectangle that defines the position of captured data on the document’s validation screen. You can always edit or re-adjust it to ensure that the annotations are correct and precise.
Line item
Any product or service listed in an invoice or any other type of document. Line item data usually includes quantity, price, and other relevant attributes.
Custom data field
It allows you to extract fields that the Generic AI Engine does not automatically capture. You must adjust the queue schema if you want to add a custom data field.
Annotation process
This process involves collecting documents and consistently highlighting the data required for extraction with precision. High-quality annotations are crucial to ensuring the AI engine learns what data to capture accurately.
Pre-trained Data Field
Data fields that Rossum’s Generic AI Engine captures automatically straight out of the box.
Check our "Rossum AI Engines" article to learn more about the currently recognized data fields.
Data Field
A data field is a single piece of information captured from a document.
You can modify a data field’s label, ID, and more for every queue. You can also hide them or delete if not needed.
For a more in-depth look at extracted field types, please visit our API documentation.
Document editing screen
This section allows you to edit documents. You can change page orientation, delete pages, and split single documents into multiple new records. To access it, click the Edit button on the validation screen.
Validation screen
A screen where users can review and confirm the data extracted from a document. It opens automatically when a document is accessed.
Manage users view
A section for Rossum account administration. Here an admin can add new users, assign them to a specific role and queue or deactivate them.
Once created, the User detail page provides more information and is an excellent place to keep track of current access management settings.
Workspace
A workspace acts as a folder where you can group related queues. It can represent a specific customer or region.
Queue
A queue is a dedicated space for importing documents that require processing.
Each queue is assigned an extraction schema - a set of fields to be extracted from the document - which can be customized based on your requirements.
Within a queue, you can configure custom automation settings and add extensions to enhance document workflows, enabling seamless integration with other systems.
AI engine
Rossum’s unique AI engines incorporate neural networks to identify and extract document data. The two main types are the Generic AI Engine and the Dedicated AI Engine. Users can use either engine to achieve the desired results based on individual needs.
Generic AI engine
It is an engine pre-trained to recognize a specific set of fields out of the box, even without user annotations.
Dedicated AI engine
It is a custom engine trained to recognize customer-provided data and is tailored to the customer’s specific data capture needs. Dedicated Engine helps process documents with diverse layouts. It can also extract custom data fields and increase levels of data capture accuracy over time.