natural-language-processing

document understanding

3260 papers • 126 benchmarks • 313 datasets

Document understanding involves document classification, layout analysis, information extraction, and DocQA.

(Image credit: Papersgraph)

Benchmarks

These leaderboards are used to track progress in document-understanding-2

Trend

Dataset

Best Model

Actions

No benchmarks available.

Libraries

i

Use these libraries to find document-understanding-2 models and implementations

huggingface/transformers

5 papers 124,061

Datasets

PDFVQA

Subtasks

No subtasks available.

Most implemented papers

LayoutXLM: Multimodal Pre-training for Multilingual Visually-rich Document Understanding

Cha Zhang, Yijuan Lu, Yiheng Xu, Tengchao Lv, Lei Cui, Guoxin Wang, D. Florêncio, Furu Wei•Sat Apr 17 2021

The LayoutXLM model has significantly outperformed the existing SOTA cross-lingual pre-trained models on the XFUND dataset and aims to bridge the language barriers for visually-rich document understanding.

167

Content

alibabaresearch/advancedliteratemac…

4 papers 678

PaddlePaddle/PaddleOCR

2 papers 38,104

hikopensource/davar-lab-ocr

2 papers 703

facebookresearch/data2vec_vision

2 papers 74

0

Paper Graph

LayoutLMv2: Multi-modal Pre-training for Visually-rich Document Understanding

Lidong Zhou, Cha Zhang, Wanxiang Che, Yijuan Lu, Yiheng Xu, Tengchao Lv, Lei Cui, Guoxin Wang, D. Florêncio, Furu Wei, Yang Xu, Min Zhang•Mon Dec 28 2020

LayoutLMv2 architecture with new pre-training tasks to model the interaction among text, layout, and image in a single multi-modal framework and achieves new state-of-the-art results on a wide variety of downstream visually-rich document understanding tasks.

621 0

Paper Graph

Chargrid: Towards Understanding 2D Documents

Anoop R. Katti, C. Reisswig, Cordula Guder, Sebastian Brarda, S. Bickel, J. Höhne, J. Faddoul•Sun Sep 23 2018

A novel type of text representation is introduced that preserves the 2D layout of a document by encoding each document page as a two-dimensional grid of characters and it is shown that it significantly outperforms approaches based on sequential text or document images.

206 0

Paper Graph

OCR-Free Document Understanding Transformer

Sangdoo Yun, Dongyoon Han, Geewook Kim, Teakgyu Hong, Moonbin Yim, JeongYeon Nam, Jinyeong Yim, Wonseok Hwang, Seunghyun Park•Mon Nov 29 2021

A novel OCR-free VDU model named Donut, which stands for Document understanding transformer, achieves state-of-the-art performances on various VD U tasks in terms of both speed and accuracy and offers a synthetic data generator that helps the model pre-training to be flexible in various languages and domains.

395 0

Paper Graph

ICDAR 2021 Competition on Scientific Literature Parsing

Xu Zhong, Antonio Jimeno-Yepes, D. Burdick•Mon Jun 07 2021

The ICDAR 2021 Scientific Literature Parsing Competition (ICDAR2021-SLP) aims to drive the advances specifically in document understanding and leverages the PubLayNet and PubTabNet datasets, which provide hundreds of thousands of training and evaluation examples.

36 0

Paper Graph

LayoutLLM: Layout Instruction Tuning with Large Language Models for Document Understanding

Chuwei Luo, Cong Yao, Yufan Shen, Zhaoqing Zhu, Qi Zheng, Zhi Yu•Sun Apr 07 2024

Experiments show that the proposed LayoutLLM significantly outperforms existing methods that adopt open-source 7B LLMs/MLLMs for document understanding, and brings a certain degree of interpretability, which could facilitate manual inspection and correction.

107 0

Paper Graph

Message Passing Attention Networks for Document Understanding

A. Tixier, M. Vazirgiannis, Giannis Nikolentzos•Fri Aug 16 2019

This paper represents documents as word co-occurrence networks and proposes an application of the message passing framework to NLP, the Message Passing Attention network for Document understanding (MPAD), and proposes several hierarchical variants of MPAD.

78 0

Paper Graph

MarkupLM: Pre-training of Text and Markup Language for Visually Rich Document Understanding

Yiheng Xu, Lei Cui, Furu Wei, Junlong Li•Fri Oct 15 2021

This paper proposes MarkupLM for document understanding tasks with markup languages as the backbone, such as HTML/XML-based documents, where text and markup information is jointly pre-trained.

68 0

Paper Graph

LiLT: A Simple yet Effective Language-Independent Layout Transformer for Structured Document Understanding

Lianwen Jin, Jiapeng Wang, Kai Ding•Sun Feb 27 2022

Experimental results on eight languages have shown that LiLT can achieve competitive or even superior performance on diverse widely-used downstream benchmarks, which enables language-independent benefit from the pre-training of document layout structure.

178 0

Paper Graph

End-to-end Document Recognition and Understanding with Dessurt

Brian L. Davis, B. Morse, Brian L. Price, Chris Tensmeyer, Vlad I. Morariu, Curtis Wigington•Tue Mar 29 2022

Dessurt is a relatively simple document understanding transformer capable of being fine-tuned on a greater variety of document tasks than prior methods and is able to handle a variety ofdocument domains and tasks.

85 0

Paper Graph

Adding a benchmark result helps the community track progress.