document

Definition

A self-contained unit of text with internal structure — a report, invoice, paper, or slide deck. It is the natural unit for retrieval, summarisation, and information extraction, and in document AI it usually means the page image and its layout as well as the characters. Distinguish it from a passage or chunk, which are the smaller pieces a document is split into before indexing.

Explainers using this term

• Document AI and OCR Today — How an LLM Ends Up Reading Your InvoicesLayoutLM: Pre-training of Text and Layout for Document Image Understanding