Structuring the Unstructured - Cedric Clyburn, Red Hat

An AI answering questions about a document needs to retain which heading, table column or caption a passage belongs to. Docling, an open-source document converter, addresses this by combining text recognition with models that identify page layout and table structure. The talk explains how this produces structured data that applications can export as Markdown or JSON.

The resulting representation keeps tables, pictures and text as separate item types within a document tree, with layout bounding boxes for items where available. Docling's document-format documentation describes this structure. That gives an application a way to retrieve a table as a table and to locate extracted material on the page. Such structure makes extraction inspectable; it does not establish that every extracted value is correct.

“I think we can all agree that context is the most important aspect to building an AI application or an agent, right?”

Cedric Clyburn

“20 scientific papers now feature a new nonsensical term that doesn't exist because AI misinterpreted a very old article that was scanned and taken to a PDF”

Cedric Clyburn

“that data and the way you process it is the key determining factor in whether your answer is going to be correct or incorrect for the user or customer at the end of the day”

Cedric Clyburn

“we're doing RAG but without having to use a chunker or embedding model or vector database, etc., etc. So the index ends up being the markdown outline of the document.”

Cedric Clyburn

“It's fast, it's cheap, and most importantly, it's open-source.”

Cedric Clyburn