Saved in:
Bibliographic Details
Main Authors: Perot, Vincent, Kang, Kai, Luisier, Florian, Su, Guolong, Sun, Xiaoyu, Boppana, Ramya Sree, Wang, Zilong, Wang, Zifeng, Mu, Jiaqi, Zhang, Hao, Lee, Chen-Yu, Hua, Nan
Format: Preprint
Published: 2023
Subjects:
Online Access:https://arxiv.org/abs/2309.10952
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913400434458624
author Perot, Vincent
Kang, Kai
Luisier, Florian
Su, Guolong
Sun, Xiaoyu
Boppana, Ramya Sree
Wang, Zilong
Wang, Zifeng
Mu, Jiaqi
Zhang, Hao
Lee, Chen-Yu
Hua, Nan
author_facet Perot, Vincent
Kang, Kai
Luisier, Florian
Su, Guolong
Sun, Xiaoyu
Boppana, Ramya Sree
Wang, Zilong
Wang, Zifeng
Mu, Jiaqi
Zhang, Hao
Lee, Chen-Yu
Hua, Nan
contents Large Language Models (LLM) have revolutionized Natural Language Processing (NLP), improving state-of-the-art and exhibiting emergent capabilities across various tasks. However, their application in extracting information from visually rich documents, which is at the core of many document processing workflows and involving the extraction of key entities from semi-structured documents, has not yet been successful. The main obstacles to adopting LLMs for this task include the absence of layout encoding within LLMs, which is critical for high quality extraction, and the lack of a grounding mechanism to localize the predicted entities within the document. In this paper, we introduce Language Model-based Document Information Extraction and Localization (LMDX), a methodology to reframe the document information extraction task for a LLM. LMDX enables extraction of singular, repeated, and hierarchical entities, both with and without training data, while providing grounding guarantees and localizing the entities within the document. Finally, we apply LMDX to the PaLM 2-S and Gemini Pro LLMs and evaluate it on VRDU and CORD benchmarks, setting a new state-of-the-art and showing how LMDX enables the creation of high quality, data-efficient parsers.
format Preprint
id arxiv_https___arxiv_org_abs_2309_10952
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle LMDX: Language Model-based Document Information Extraction and Localization
Perot, Vincent
Kang, Kai
Luisier, Florian
Su, Guolong
Sun, Xiaoyu
Boppana, Ramya Sree
Wang, Zilong
Wang, Zifeng
Mu, Jiaqi
Zhang, Hao
Lee, Chen-Yu
Hua, Nan
Computation and Language
Artificial Intelligence
Machine Learning
Large Language Models (LLM) have revolutionized Natural Language Processing (NLP), improving state-of-the-art and exhibiting emergent capabilities across various tasks. However, their application in extracting information from visually rich documents, which is at the core of many document processing workflows and involving the extraction of key entities from semi-structured documents, has not yet been successful. The main obstacles to adopting LLMs for this task include the absence of layout encoding within LLMs, which is critical for high quality extraction, and the lack of a grounding mechanism to localize the predicted entities within the document. In this paper, we introduce Language Model-based Document Information Extraction and Localization (LMDX), a methodology to reframe the document information extraction task for a LLM. LMDX enables extraction of singular, repeated, and hierarchical entities, both with and without training data, while providing grounding guarantees and localizing the entities within the document. Finally, we apply LMDX to the PaLM 2-S and Gemini Pro LLMs and evaluate it on VRDU and CORD benchmarks, setting a new state-of-the-art and showing how LMDX enables the creation of high quality, data-efficient parsers.
title LMDX: Language Model-based Document Information Extraction and Localization
topic Computation and Language
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2309.10952