AI Paper Atlas

Research map / Graph learning / Knowledge and temporal graphs

DocTr: Document Transformer for Structured Information Extraction in Documents

Haofu Liao, Aruni RoyChowdhury, Weijian Li, Ankan Bansal, Yuting Zhang, Zhuowen Tu, Ravi Kumar Satzoda, R. Manmatha, Vijay Mahadevan

ICCV 2023 · Graph learning › Knowledge and temporal graphs · approach: Graph-based dialog systems · 43 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
use anchor word detection formulation instead of IOB tagging (raster scan text) (ablation)DocTr backbone · CORD receipt parsingF1: 93.2 -> 94.4Improved
use anchor word detection formulation instead of SPADE graph decoding (ablation)DocTr backbone · CORD receipt parsingF1: 93.0 -> 94.4Improved
switch from oracle to raster scan text order (ablation)anchor word formulation vs IOB/SPADE · CORD receipt parsingF1 drop: 0.6 drop vs 0.9 dropImproved
fine-tune general-purpose models with the anchor word formulation instead of IOB taggingLayoutLMv3 · CORD receipt parsingF1: 92.2 -> 93.6Improved
use first and last words as anchors instead of first word only (ablation)DocTr · CORD receipt parsingF1: 94.2 -> 94.4Improved
use last word instead of first word as anchor (ablation)DocTr · CORD receipt parsingF1: 94.2 -> 94.1No clear effect
use name anchor as primary anchor instead of first entity of line item (ablation)DocTr · CORD receipt parsingF1: 94.0 -> 94.4Improved
add masked detection modeling pre-training instead of none (ablation)DocTr · CORD receipt parsingF1: 82.3 -> 94.4Improved
use MDM instead of MVLM pre-training (ablation)DocTr · CORD receipt parsingF1: 90.9 -> 94.4Improved
use MDM instead of MVLM pre-training (ablation)DocTr · FUNSD entity linkingF1: 73.0 -> 73.9Improved
add vision encoder and VL-decoder to language encoder (ablation)DocTr · CORD receipt parsingF1: 92.3 -> 94.4Improved
add VL-decoder without vision encoder (ablation)DocTr · CORD receipt parsingF1: 92.3 -> 92.6No clear effect

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