UiPath® Helix Extractor 1.0

The Helix Extractor 1.0 large language model (LLM) is our latest data extraction model technology, designed to replace current generation models used within UiPath® Document Understanding^TM. While Helix Extractor 1.0 operates similarly to previous models, it was trained using a wide variety of documents. This enables it to process common document types with little to no training needed. What sets Helix Extractor 1.0 LLM apart is its generative architecture, which significantly improves accuracy and simplifies extraction. Additionally, you can also fine-tune the model with your unique datasets.

To gain further insights into the Helix Extractor 1.0 architecture and the techniques used for training, check the Helix Extractor 1.0 page from our AI blog.

Disponibilidad

Currently, the UiPath Helix Extractor is only available for US-based tenants (excluding GxP and Government Cloud) in Document Understanding modern projects.

The UiPath Helix Extractor is available for both classic and modern projects when using public endpoints in the following regions:

• Public endpoints for extraction models in Europe are based on the Helix Extractor, except for Financial Statements.
• The following public endpoints for extraction models are based on the Helix Extractor in the Japan region:
  • Facturas China
  • Facturas Japón
  • Recibos de Japón

Mejoras con respecto a la generación anterior

The Helix Extractor LLM offers numerous enhancements over previous models. It improves accuracy, especially with tables, adapts to various document layouts to reduce annotation efforts, and boosts automation rates.

Las mejoras clave incluyen:

• Improved accuracy: The Helix Extractor LLM delivers a higher accuracy rate and superior F1 score for semi-structured documents such as invoices, receipts, and purchase orders. This ensures precise and consistent data extraction.
• Anotación sin esfuerzo: el modelo reduce el trabajo manual al requerir solo una anotación por documento, eliminando la necesidad de anotar cada instancia de campo en cada página.
• Enhanced automation: With a greater correlation between confidence level and accuracy, the Helix Extractor LLM enhances automation rates while reducing the number of documents sent to Action Center for the same accuracy level.

From our internal tests, the Helix Extractor outperformed its predecessor in performance. It reduced the false positive rate by around 15%, and the false negative rate dropped by nearly 17%.

How to use the Helix Extractor

The Helix Extractor LLM is available exclusively for Document Understanding modern projects. Despite the introduction of the Helix Extractor, all existing project versions will still use current model versions. This ensures a seamless transition without any disruption to ongoing production workflows.

To start training an exisiting document type on the Helix Extractor, unconfirm and confirm all fields in a few documents.

1. Choose the document type you want to train on the Helix Extractor.
2. Selecciona un documento.
3. Selecciona todos los campos del documento y elige Eliminar.
   
   

   

   
   
4. Anota todos los campos del documento y selecciona Confirmar.
   Nota: repite los pasos 3 y 4 hasta que se inicie el entrenamiento en el tipo de documento elegido.

   
   

   

   
   

How to check if the Helix Extractor is enabled

After training your models on the Helix Extractor, check the model version to make sure that the Helix Extractor is enabled.

1. Ve a la página Publicar y crea una nueva versión del proyecto.
2. Selecciona el icono de tres puntos ⋮ junto a la versión del proyecto y elige Editar versión para comprobar la versión del modelo.
   Note: All models version 24.7 and above are UiPath Helix Extractor models.

   
   

   

   
   

Optimización de resultados

Los nombres de campo que elijas pueden afectar en gran medida al rendimiento del modelo. Para garantizar resultados óptimos, utiliza el lenguaje natural y la gramática adecuada para los nombres de campo. Solo debes utilizar acrónimos ampliamente reconocidos como Número (No), Cuenta (Acct), Dirección (Addr) y Apartamento (Apt). Actualmente, solo se admiten idiomas de Europa occidental, así que asegúrate de que los nombres de campo elegidos se alineen con estos idiomas. Evita utilizar nombres no descriptivos, como "Columna 3", a menos que el documento utilice específicamente esa terminología.

Choosing between the Helix Extractor and legacy model type

The UiPath Helix Extractor currently supports only Latin script languages. If you need to train a model in non-Latin script languages, choose the legacy model type. If the legacy model is selected, choose the appropriate base model for your document type.

To choose between the Helix Extractor or legacy model type, navigate to the Settings tab in Document Type Manager and select the needed model type from the Model type drop-down list.

Importante: Es necesario publicar una nueva versión del proyecto después de implementar los cambios.

UiPath® Helix Extractor known limitations

The following limitations currently apply for UiPath Helix Extractor:

• Los campos extraídos deben coincidir exactamente con el texto de los documentos. Este proceso no incluye resumir u otros tipos de análisis de texto.
• The following document types are not currently based on the Helix Extractor and still work on the previous generation:
  • Estados financieros
  • Facturas China
  • Facturas en hebreo
  • Facturas Japón

Tip: Document types that are not currently supported by the Helix Extractor model have the following message in the Add document type drop-down list:

El tipo de documento se entrenará utilizando el modelo heredado.

The UiPath Helix Extractor does not currently support non-Latin script languages.