Document Understanding
latest
false
Banner background image
PREVIEW
Document Understanding User Guide for Modern Experience
Last updated Feb 22, 2024

Protection of Personal Data and Privacy of the Users

Document Understanding is an AI product which brings the most advanced and powerful Deep Learning and Generative AI technologies to customers' fingertips. This is one of its key characteristics, and depends on data availability. Without data, the product would not exist and would not be able to be maintained and operating within the Service Level Agreements committed to customers.

The product faces several challenges (availability SLAs, reliability, scalability, performance and accuracy), and data availability is required for all of them.

Certain kinds of data inputs can cause services to become unavailable, and maintaining service SLAs and responding to issues requires access to the root cause, which is usually the document payload. The most common examples are documents with many pages, documents with very dense text on each page,documents with certain special encodings, fonts, languages, or file formats especially for PDFs, TIFFs and PNGs.

Service performance, especially the latency may also be impacted by very specific features of the input data, and there are many attended scenarios where response time is critical for having a viable product.

Finally, we can only monitor and ensure sustained high performance by accessing realistic documents, for both testing and training. For instance, as a new model architecture becomes available and we want to provide its benefits to our customers, it may happen that the overall performance improves, but for specific types of input, the performance degrades, sometimes severely. There is no way for us to detect and prevent this degradation without access to a broad, diverse and comprehensive set of real-world data.

In short, data forms an integral part of the product development, maintenance and support cycles for Document Understanding.
  • Product limitation. The Document Understanding team confirms that data collected from Document Understanding can only be used to support, maintain and improve the Document Understanding product.
  • Deletion of data. The Document Understanding team confirms that the data pertaining to each customer can be kept for retraining
    • Until the customer requests its deletion, or
    • If the contractual arrangement requires to delete the data at the end of the relationship.

Was this page helpful?

Get The Help You Need
Learning RPA - Automation Courses
UiPath Community Forum
Uipath Logo White
Trust and Security
© 2005-2024 UiPath. All rights reserved.