Uploaded image for project: 'Red Hat Enterprise Linux AI'
  1. Red Hat Enterprise Linux AI
  2. RHELAI-3853

[UX] Chunking & Assessment

XMLWordPrintable

    • Icon: Story Story
    • Resolution: Done
    • Icon: Undefined Undefined
    • None
    • None
    • Fit and Finish
    • False
    • Hide

      None

      Show
      None
    • False

      Problem Statement

      When Docling ingests documents, there is a chunking strategy used. However, the user may want to modify the way their document has been chunked. For example, this may be to remove unnecessary content or to change the bounding boxes. The user needs the proper guidance based on the goal of that knowledge contribution.

      Goal

      Define the experience for editing the documents chunks and properly guiding the user to do so.

      Open Questions

      • How could policies help with reducing the cognitive load for the user?
      • How can we provide an experience that scales?
      • How can we help them scan to identify what they do and do not care about when something does not convert properly?
      • How can we guide users to identify what content gets into the model?

      Acceptance Criteria

      • We need to identify and undertand where will users need help - documenting running these tools/where they will be run, Jupyter nbs etc.
        • How can we make it better?

      Resources

      Recording from the FnF demo on April 3 - [link to come], where we talk about separating the content experience from the authoring of QnA. Mo provides high level explanation of the why.

      Deliverables

      Miro

              mdenisco Maria Denisco
              beaumorley Beau Morley
              Votes:
              0 Vote for this issue
              Watchers:
              1 Start watching this issue

                Created:
                Updated:
                Resolved: