On a new field, the Identification Supervision setting will initially be set to Default and can subsequently be changed by the user. Always - a field with this setting will always generate Field ID tasks, regardless of the machine’s confidence in identifying this field. This setting is appropriate for scenarios where you want to guarantee a manual review of a field. Notes (optional) - setting allows notes to be added to a field’s definition. Adding new fields to a live layout causes a warning message to appear on the layout’s details page. This warning message prompts you to go to the Release Library ( Library > Releases ) and create a new release with your updated layout variation. If you add new fields to a layout that has a live Field Locator model, a warning message informs you that the ground-truth data does not include the newly-added fields. Additional …

If you want to use Hyperscience to automate the processing of submissions in multiple languages, keep the following details in mind during your planning efforts.  For more information on the languages Hyperscience supports, see Supported Languages . Multiple languages within a submission When a submission is ingested into Hyperscience, the system matches its pages to one or more layout variations for automation. Each set of pages matched to a layout variation becomes a document , and each layout variation has a single language assigned to it. Therefore, each document within a submission can be automated in a single language, but each submission can include documents in multiple languages (e.g., two French documents and one English document). To learn more about creating layouts, see Creating Structured Layouts , Creating Semi-structured Layouts , and Creating Additional Layouts . Using a layout across languages Hyperscience does not support the use of duplicate layouts in an …

Overview With the Message Queue Listener Input Block, you can configure connections to the following message queue types: ActiveMQ Amazon SQS IBM MQ RabbitMQ To select the message queue type you want to use for a Message Queue Listener connection, click on its name in the Message Queue Type drop-down list at the top of the block’s settings. If you would like to use a Message Queue Listener connection in a SaaS instance and your MQ service is not public, you should create a firewall pinhole for the connection. Contact your Hyperscience representative for more information. Sample use cases I send submission info from external systems to an Amazon SQS queue. I want to ingest this submission info into Hyperscience for processing. My customers will post relevant documents and metadata to my ActiveMQ queue. I want to then ingest these documents and metadata for processing into Hyperscience …

tudio project, including the XAML definition file, is attached. Import the XAML file in UiPath Studio as a UiPath Activity. The Activity can then be added in larger workflows and set up to process notifications on Hyperscience document submission state changes received on a UiPath Queue. Finally, the Activity should be configured to forward the notification data for further processing depending on your business rules and processes. The following occurs when a UiPath robot executes this workflow: If the queue is empty, the robot waits some time (e.g. 10 seconds) before checking the queue again. If the queue is not empty, a notification is retrieved, acknowledged, and sent for further processing. The contents of the notification are saved in a variable called notification , which is in the format of a Hyperscience submission object as it has been defined in our API documentation . The Flow Switch routes the UiPath robot …

...

Each area of the application can be accessed by clicking on it in the left-hand sidebar or in the navigation bar across the top of the application. These options will be shown on all application screens, except for the Supervision and QA task views, the Layout Editor, Flow Studio, and the Release Editor. Below is a list and summary of each part of the application, along with the types of users who would typically use that specific area. The interface Navigation section Section purpose Users Perform Tasks tab ( Tasks > Perform Tasks , or click on the Hyperscience logo) Provides keyer access to manual tasks awaiting completion. System Admin, Business Admin, Data Keyer Admin, Data Keyer, Knowledge Worker Tasks Overview tab ( Tasks > Overview ) Displays metrics on in-progress and recently completed work. System Admin, Business Admin, Data Keyer Admin, Data Keyer, Knowledge Worker Task Queue  ( Tasks > Task Queue ) Lists outstanding …

...

asks. Collation Block We’ve created a Collation Block to allow grouping of files, documents, and pages into cases. To learn more about Case Collation, see Case Collation . Settings You can customize the functionality of your block by editing the settings described below. Name Required? Description Replace case data from duplicate file names No Replaces case data from repeated file names within the same case. For example, you can enable this setting if you want to resubmit a file containing new data and remove the old, duplicated data from the case. Note that this setting does not delete the old data; it just removes it from the case.  Custom Supervision Block To enable the tailoring of a Supervision task’s interface to a specific business process, we’ve created a Custom Supervision Block. To use Custom Supervision, you need a custom flow with a Custom Code Block: A custom flow …

While there are many ways to customize individual flows, you may also want to configure custom behavior outside of a flow for submissions processed through certain blocks. For example, you may need to send submissions' state-change information to your downstream systems mid-flow. With the Notification Flow field in Submission Initialization and Manual Supervision Blocks’ settings, you can link Notification flows to specific blocks in your Document Processing flow. Linking from pre-built vs. custom flows If you are linking from one of the pre-built flows in your instance (e.g., "Document Processing V36"), you can only link to Notification flows. However, you can create customized blocks with the Flows SDK that allow you to link to any flow in your instance, regardless of whether it's a Notification flow or a Document Processing flow. The examples in this article show Document Processing flows linking to Notification …

ing to reach the target accuracy set in Structured Signature Accuracy . The system automatically calculates this value after the minimum amount of training data is obtained through Transcription QA. N/A Structured Signature Threshold This setting determines the minimum confidence thresholds needed for a signature to be automatically processed. Signatures with confidence scores above this threshold are automatically processed. If a signature’s confidence score is below this threshold, it will be sent to Transcription Supervision. If Transcription Automation Training for Structured documents is enabled, any value you enter manually will be overwritten by the value calculated by the system, based on your target accuracy. .56 Structured Signature Minimum Legibility Threshold The minimum confidence score a signature field must have in a Structured document in order for the system to automatically process the field. If a signature field's confidence score is below this value, the system will mark it …

The Automation Training Result chart reports the combined projected automation for all language families with and without automation training, based on the current target accuracy. To learn more about language families, see Supported Languages . To see more specific metrics related to each data point, hover your cursor over a data point of interest and a window will display the overall accuracy as well as automation percentages with and without training.  

w these steps: Click  Download Related Pages  from the "Download" dropdown button and extract the files from the downloaded ZIP file. Open the CSV which includes information about each related image. This information is useful in identifying which images are most likely a part of your multi-page document. Creating layout variations from job results Keep in mind that downloaded images may still contain PII or other sensitive information and may require manual redaction – review the images carefully prior to downloading, utilize the Manual Redaction feature if necessary, and ensure that any downloaded images are handled according to applicable data-retention policies.

In v36, models for flows created in v35 are forward compatible , meaning that you can use them in v36 without having to retrain them during the upgrade process. As a result, forward-compatible models allow you to: upgrade v35 models after upgrading the application to v36, maintain the automation levels achieved in v35, and continue existing training efforts for any use cases that may be in implementation when upgrading to v36. Which models are forward compatible In v36, the following types of models are forward compatible: Locator models Transcription models (i.e., sets of finetuning models) Classification models The following types of models are not forward compatible, so you need to train v36 versions of these models in order to use them in v36: Text Classification models To learn more about Text Classification, see Text Classification . Forward-compatible models and upgrades When upgrading to v36, keep in mind that …

Overview Similar to Field ID QA, the machine measures transcription accuracy by establishing Quality Assurance (QA) consensus; consensus means that two entries for a given field must have matching normalized transcriptions. In other words, when the transcription of the machine (with high confidence) or that of a data keyer matches what is determined in QA consensus, then the data keyer or the machine is marked as accurate.  With Release 25.0.0, the system scores every human entry for Field ID. However, as in previous releases, there are specific instances where this is not the case: The machine is only scored when the confidence is above the Transcription threshold. If the confidence is below the threshold/field was sent to Supervision, the machine is not scored for accuracy. However, the response may be used to establish consensus and measure human accuracy. When a field requires human intervention during processing …

Documents can sometimes have layouts that look almost identical, with only a small difference in their visual templates. As a best practice, we recommend creating layout variations to account for these minor differences, as described in Adding a Variation to a Layout . Even with layout variations, the system may classify documents incorrectly. To prevent that from happening, we’ve introduced settings that help the system improve classification accuracy by using indications from the document’s pages after the system’s initial matching. These settings force the system to match documents based specifically on the layout identifier. If the layout identifier in the document matches the expected ID in the layout variation, the system will classify the document according to that layout variation. This matching is valid regardless of whether the initial confidence score was below or above the threshold. The system is effectively forcing the match, ensuring accuracy. If …

Overview With the Message Queue Notifier Output Block, you can configure connections to the following message queue types: ActiveMQ Amazon SQS IBM MQ RabbitMQ To select the message queue type you want to use for a Message Queue Notifier connection, click on its name in the Message Queue Type drop-down list at the top of the block’s settings. If you would like to use a Message Queue Notifier connection in a SaaS instance and your MQ service is not public, you should create a firewall pinhole for the connection. Contact your Hyperscience representative for more information. Sample use cases I have extracted content from insurance forms and want to send the JSON downstream to my IBM MQ in order to populate other systems. I have transformed submission data and want to send it to my RabbitMQ, where it will be picked up by internal software to be …

rows, and review the resultant grid before sending the table for Transcription. Table ID model – Train a machine learning model to automatically identify the table column fields. Model validation tasks – Boost model performance with a Supervision task that asks a human to correct errors. Tables reporting – View an automation report that shows how many columns the machine predicted correctly, a throughput report showing how much table data is being processed, and a volume report showing how much data is passing through Supervision tasks.  Notes about this feature: Table ID automation only applies to documents that contain one table and follow a standard grid format. A standard grid format refers to tables where the data falls neatly within the boundaries of each cell, as defined by the rows and columns of the table, and the information contained has a 1:1 relationship with a corresponding column and row.  Removed the “Automation …

...

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 below the transcription target accuracy thresholds. For more information on how to set thresholds, see Transcription Settings . If a field has Consensus or the "Supervision Always" setting enabled in the Layout Editor. Note that these tasks have escalated priority. Flexible Extraction tasks are generated in the following situations: A Structured page was originally marked as No Layout Found by the machine, but it was later manually matched to a layout in Document Classification Supervision. The code in a Custom Code Block sent a document to Flexible Extraction so a keyer could verify the content of specific fields. The code in a Custom Code Block sent a document to Flexible Extraction so a keyer could …