TWIX: Automatically Reconstructing Structured Data from Templatized Documents

• URL: http://arxiv.org/abs/2501.06659v1
• Date: Sat, 11 Jan 2025 23:07:04 GMT
• Title: TWIX: Automatically Reconstructing Structured Data from Templatized Documents
• Authors: Yiming Lin, Mawil Hasan, Rohan Kosalge, Alvin Cheung, Aditya G. Parameswaran,
• Abstract summary: Our tool TWIX predicts the underlying template used to create templatized documents.<n>TWIX achieves over 90% precision and recall on average, outperforming tools from industry.<n> TWIX scales easily to large datasets and is 734X faster and 5836X cheaper than vision-based LLMs for extracting data from a large document collection with 817 pages.
• Score: 11.03654616939188
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Many documents, that we call templatized documents, are programmatically generated by populating fields in a visual template. Effective data extraction from these documents is crucial to supporting downstream analytical tasks. Current data extraction tools often struggle with complex document layouts, incur high latency and/or cost on large datasets, and often require significant human effort, when extracting tables or values given user-specified fields from documents. The key insight of our tool, TWIX, is to predict the underlying template used to create such documents, modeling the visual and structural commonalities across documents. Data extraction based on this predicted template provides a more principled, accurate, and efficient solution at a low cost. Comprehensive evaluations on 34 diverse real-world datasets show that uncovering the template is crucial for data extraction from templatized documents. TWIX achieves over 90% precision and recall on average, outperforming tools from industry: Textract and Azure Document Intelligence, and vision-based LLMs like GPT-4-Vision, by over 25% in precision and recall. TWIX scales easily to large datasets and is 734X faster and 5836X cheaper than vision-based LLMs for extracting data from a large document collection with 817 pages.

Related papers

• BiblioPage: A Dataset of Scanned Title Pages for Bibliographic Metadata Extraction [0.0]
  BiblioPage is a dataset of scanned title pages annotated with structured metadata. The dataset consists of approximately 2,000 title pages collected from 14 Czech libraries. We valuated object detection models such as YOLO and DETR combined with transformer-based OCR, achieving a maximum mAP of 52 and an F1 score of 59.
  arXiv  Detail & Related papers  (2025-03-25T13:46:55Z)
• BigDocs: An Open Dataset for Training Multimodal Models on Document and Code Tasks [57.589795399265945]
  We introduce BigDocs-7.5M, a high-quality, open-access dataset comprising 7.5 million multimodal documents across 30 tasks. We also introduce BigDocs-Bench, a benchmark suite with 10 novel tasks. Our experiments show that training with BigDocs-Bench improves average performance up to 25.8% over closed-source GPT-4o.
  arXiv  Detail & Related papers  (2024-12-05T21:41:20Z)
• Document Classification using File Names [7.130525292849283]
  Rapid document classification is critical in several time-sensitive applications like digital forensics and large-scale media classification. Traditional approaches that rely on heavy-duty deep learning models fall short due to high inference times over vast input datasets and computational resources associated with analyzing whole documents. We present a method using lightweight supervised learning models, combined with a TF-IDF feature extraction-based tokenization method, to accurately and efficiently classify documents based solely on file names.
  arXiv  Detail & Related papers  (2024-10-02T01:42:19Z)
• CRAFT Your Dataset: Task-Specific Synthetic Dataset Generation Through Corpus Retrieval and Augmentation [51.2289822267563]
  We propose Corpus Retrieval and Augmentation for Fine-Tuning (CRAFT), a method for generating synthetic datasets. We use large-scale public web-crawled corpora and similarity-based document retrieval to find other relevant human-written documents. We demonstrate that CRAFT can efficiently generate large-scale task-specific training datasets for four diverse tasks.
  arXiv  Detail & Related papers  (2024-09-03T17:54:40Z)
• SynthDoc: Bilingual Documents Synthesis for Visual Document Understanding [23.910783272007407]
  This paper introduces SynthDoc, a novel synthetic document generation pipeline designed to enhance Visual Document Understanding (VDU) Addressing the challenges of data acquisition and the limitations of existing datasets, SynthDoc leverages publicly available corpora and advanced rendering tools to create a comprehensive and versatile dataset. Our experiments, conducted using the Donut model, demonstrate that models trained with SynthDoc's data achieve superior performance in pre-training read tasks and maintain robustness in downstream tasks, despite language inconsistencies.
  arXiv  Detail & Related papers  (2024-08-27T03:31:24Z)
• DocGenome: An Open Large-scale Scientific Document Benchmark for Training and Testing Multi-modal Large Language Models [63.466265039007816]
  We present DocGenome, a structured document benchmark constructed by annotating 500K scientific documents from 153 disciplines in the arXiv open-access community. We conduct extensive experiments to demonstrate the advantages of DocGenome and objectively evaluate the performance of large models on our benchmark.
  arXiv  Detail & Related papers  (2024-06-17T15:13:52Z)
• Enhancing Document Information Analysis with Multi-Task Pre-training: A Robust Approach for Information Extraction in Visually-Rich Documents [8.49076413640561]
  The model is pre-trained and subsequently fine-tuned for various document image analysis tasks. The proposed model achieved impressive results across all tasks, with an accuracy of 95.87% on the RVL-CDIP dataset for document classification.
  arXiv  Detail & Related papers  (2023-10-25T10:22:30Z)
• DocumentNet: Bridging the Data Gap in Document Pre-Training [78.01647768018485]
  We propose a method to collect massive-scale and weakly labeled data from the web to benefit the training of VDER models. The collected dataset, named DocumentNet, does not depend on specific document types or entity sets. Experiments on a set of broadly adopted VDER tasks show significant improvements when DocumentNet is incorporated into the pre-training.
  arXiv  Detail & Related papers  (2023-06-15T08:21:15Z)
• DocILE Benchmark for Document Information Localization and Extraction [7.944448547470927]
  This paper introduces the DocILE benchmark with the largest dataset of business documents for the tasks of Key Information Localization and Extraction and Line Item Recognition. It contains 6.7k annotated business documents, 100k synthetically generated documents, and nearly1M unlabeled documents for unsupervised pre-training.
  arXiv  Detail & Related papers  (2023-02-11T11:32:10Z)
• Layout-Aware Information Extraction for Document-Grounded Dialogue: Dataset, Method and Demonstration [75.47708732473586]
  We propose a layout-aware document-level Information Extraction dataset, LIE, to facilitate the study of extracting both structural and semantic knowledge from visually rich documents. LIE contains 62k annotations of three extraction tasks from 4,061 pages in product and official documents. Empirical results show that layout is critical for VRD-based extraction, and system demonstration also verifies that the extracted knowledge can help locate the answers that users care about.
  arXiv  Detail & Related papers  (2022-07-14T07:59:45Z)
• AQuaMuSe: Automatically Generating Datasets for Query-Based Multi-Document Summarization [17.098075160558576]
  We propose a scalable approach called AQuaMuSe to automatically mine qMDS examples from question answering datasets and large document corpora. We publicly release a specific instance of an AQuaMuSe dataset with 5,519 query-based summaries, each associated with an average of 6 input documents selected from an index of 355M documents from Common Crawl.
  arXiv  Detail & Related papers  (2020-10-23T22:38:18Z)
• SciREX: A Challenge Dataset for Document-Level Information Extraction [56.83748634747753]
  It is challenging to create a large-scale information extraction dataset at the document level. We introduce SciREX, a document level IE dataset that encompasses multiple IE tasks. We develop a neural model as a strong baseline that extends previous state-of-the-art IE models to document-level IE.
  arXiv  Detail & Related papers  (2020-05-01T17:30:10Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.