AI Paper Atlas

Research map / Vision-language and multimodal / Human and embodied modeling

American Stories: A Large-Scale Structured Text Dataset of Historical U.S. Newspapers

Melissa Dell, Jacob Carlson, Tom Bryan, Emily Silcock, Abhishek Arora, Zejiang Shen, Luca D'Amico-Wong, Quan Le, Pablo Querubin, Leander Heldring

NeurIPS 2023 · Vision-language and multimodal › Human and embodied modeling · approach: Layout-aware document understanding · 7 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
apply lightweight spellchecker after OCRAmerican Stories pipeline · ten labeled scansCER: .051 -> .044Improved
apply spellchecker to headlinesAmerican Stories pipeline · headlinesCER: .089 -> .092Hurt
use neural classifier instead of keyword methodstopic classification · article levelF1: 58.1 -> 83.6Improved
use structured American Stories instead of Chronicling America page OCR for neural classificationRoBERTa classifier · page level topicF1: 83.3 -> 96.0Improved
use American Stories for sparse mining at page levelkeyword mining · page level topicF1: 83.7 -> 82.8No clear effect
use neural method instead of Viral Texts at page levelreproduced content detection · page levelARI: 74.6 -> 86.2Improved
apply Viral Texts to American Stories instead of Chronicling AmericaViral Texts · page levelARI: 71.4 -> 74.6Improved
use EffOCR-Word (Small) instead of EffOCR-C (Base)EffOCR · Chronicling America textlinesCER: 0.023 -> 0.015Improved
use EffOCR-Word instead of TesseractOCR engine · Chronicling America textlinesCER: 0.106 -> 0.015Improved
use EffOCR-Word instead of Google Cloud VisionOCR engine · Chronicling America textlinesCER: 0.005 -> 0.015Hurt
fine-tune TrOCR (Base) on Chronicling AmericaTrOCR (Base) · day-per-decadeCER: 0.038 -> 0.027Improved
train TrOCR (Base) from scratch on target dataTrOCR (Base) · Chronicling America textlinesCER: 0.015 -> 0.809Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy