DocTr: Document Transformer for Structured Information Extraction in Documents
This work addresses the problem of extracting structured information from documents for applications like data processing, offering a novel method that improves robustness and performance.
The paper tackles structured information extraction from visually rich documents by proposing a new formulation that represents entities as anchor words and bounding boxes, addressing limitations of existing methods. The approach, including a Document Transformer and pre-training strategy, outperforms existing solutions on three benchmarks.
We present a new formulation for structured information extraction (SIE) from visually rich documents. It aims to address the limitations of existing IOB tagging or graph-based formulations, which are either overly reliant on the correct ordering of input text or struggle with decoding a complex graph. Instead, motivated by anchor-based object detectors in vision, we represent an entity as an anchor word and a bounding box, and represent entity linking as the association between anchor words. This is more robust to text ordering, and maintains a compact graph for entity linking. The formulation motivates us to introduce 1) a DOCument TRansformer (DocTr) that aims at detecting and associating entity bounding boxes in visually rich documents, and 2) a simple pre-training strategy that helps learn entity detection in the context of language. Evaluations on three SIE benchmarks show the effectiveness of the proposed formulation, and the overall approach outperforms existing solutions.