mlsanityData to 7 Oct 2026

Radically Lower Data-Labeling Costs for Visually Rich Document Extraction Models

arXiv 2210.16391 ↗First posted 28 Oct 2022cs.CLcited by 0

Not on the map. Why?

Abstract (arXiv)

A key bottleneck in building automatic extraction models for visually rich documents like invoices is the cost of acquiring the several thousand high-quality labeled documents that are needed to train a model with acceptable accuracy. We propose Selective Labeling to simplify the labeling task to provide "yes/no" labels for candidate extractions predicted by a model trained on partially labeled documents. We combine this with a custom active learning strategy to find the predictions that the model is most uncertain about. We show through experiments on document types drawn from 3 different domains that selective labeling can reduce the cost of acquiring labeled data by 10× with a negligible loss in accuracy.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 24 of its 51 references that are in the map. First posted on arXiv 28 Oct 2022.

Builds on

24 of 51 references.

References · 24

  1. 2017On Calibration of Modern Neural Networkscited by 4,860
  2. 2015Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learningcited by 4,416
  3. 2019On the Variance of the Adaptive Learning Rate and Beyondcited by 763
  4. 2019Deep Batch Active Learning by Diverse, Uncertain Gradient Lower Boundscited by 602
  5. 2019LayoutLM: Pre-training of Text and Layout for Document Image Understandingcited by 549
  6. 2020A Survey of Deep Active Learningcited by 547
  7. 2019BatchBALD: Efficient and Diverse Batch Acquisition for Deep Bayesian Active Learningcited by 428
  8. 2020LayoutLMv2: Multi-modal Pre-training for Visually-Rich Document Understandingcited by 406
  9. 2017Snorkel: Rapid Training Data Creation with Weak Supervisioncited by 370
  10. 2017Suggestive Annotation: A Deep Active Learning Framework for Biomedical Image Segmentationcited by 145
  11. 2020Cold-start Active Learning through Self-supervised Language Modelingcited by 128
  12. 2019Consistency-based Semi-supervised Active Learning: Towards Minimizing Labeling Costcited by 114
12 more
  1. 2019Discriminative Active Learningcited by 97
  2. 2019Deep Active Learning: Unified and Principled Method for Query and Trainingcited by 92
  3. 2018Deep Bayesian Active Learning for Natural Language Processing: Results of a Large-Scale Empirical Studycited by 90
  4. 2022A Survey on Programmatic Weak Supervisioncited by 66
  5. 2022FormNet: Structural Encoding beyond Sequential Modeling in Form Document Information Extractioncited by 64
  6. 2020LAMBERT: Layout-Aware (Language) Modeling for information extractioncited by 53
  7. 2019Sampling Bias in Deep Active Classification: An Empirical Studycited by 36
  8. 2021Cartography Active Learningcited by 26
  9. 2021Form2Seq : A Framework for Higher-Order Form Structure Extractioncited by 16
  10. 2017Fonduer: Knowledge Base Construction from Richly Formatted Datacited by 13
  11. 2021When Deep Learners Change Their Mind: Learning Dynamics for Active Learningcited by 6
  12. 2020Transfer Learning for Information Extraction with Limited Datacited by 4

Built on by

Not cited yet.