Attend, copy, parse end-to-end information extraction from documents

Rasmus Berg Palm, Florian Laws, Ole Winther

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

Abstract

Document information extraction tasks performed by humans create data consisting of a PDF or document image input, and extracted string outputs. This end-to-end data is naturally consumed and produced when performing the task because it is valuable in and of itself. It is naturally available, at no additional cost. Unfortunately, state-of-the-art word classification methods for information extraction cannot use this data, instead requiring word-level labels which are expensive to create and consequently not available for many real life tasks. In this paper we propose the Attend, Copy, Parse architecture, a deep neural network model that can be trained directly on end-to-end data, bypassing the need for word-level labels. We evaluate the proposed architecture on a large diverse set of invoices, and outperform a state-of-the-art production system based on word classification. We believe our proposed architecture can be used on many real life information extraction tasks where word classification cannot be used due to a lack of the required word-level labels.
Original languageEnglish
Title of host publicationProceedings of 2019 International Conference on Document Analysis and Recognition
PublisherIEEE
Publication date2019
Pages329-336
Article number8977951
ISBN (Print)9781728128610
DOIs
Publication statusPublished - 2019
Event15th IAPR International Conference on Document Analysis and Recognition - International Convention Centre, Sydney, Australia
Duration: 20 Sep 201925 Sep 2019

Conference

Conference15th IAPR International Conference on Document Analysis and Recognition
LocationInternational Convention Centre
CountryAustralia
CitySydney
Period20/09/201925/09/2019

Cite this

Palm, R. B., Laws, F., & Winther, O. (2019). Attend, copy, parse end-to-end information extraction from documents. In Proceedings of 2019 International Conference on Document Analysis and Recognition (pp. 329-336). [8977951] IEEE. https://doi.org/10.1109/icdar.2019.00060