AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Evaluating Out-of-Distribution Performance on Document Image Classifiers

Stefan Larson, Gordon Lim, Yutong Ai, David Kuang, Kevin Leach

NeurIPS 2022 · Label-efficient and robust learning › Recognition and long-tail · approach: Learning from unlabeled and positive-unlabeled data · 3 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
evaluate on RVL-CDIP-N instead of RVL-CDIP testDiT-base · RVL-CDIP vs RVL-CDIP-Nmicro accuracy: 0.933 -> 0.786Hurt
evaluate on RVL-CDIP-N instead of RVL-CDIP testVGG-16 · RVL-CDIP vs RVL-CDIP-Nmicro accuracy: 0.905 -> 0.668Hurt
evaluate on RVL-CDIP-N instead of RVL-CDIP testLayoutLMv2 · RVL-CDIP vs RVL-CDIP-Nmicro accuracy: 0.887 -> 0.556Hurt
evaluate on RVL-CDIP-N instead of RVL-CDIP testAlexNet · RVL-CDIP vs RVL-CDIP-Nmicro accuracy: 0.885 -> 0.567Hurt
add Augraphy scanner noise to RVL-CDIP-N test images (ablation)DiT-base · RVL-CDIP-Nmicro accuracy: 0.786 -> 0.706Hurt
add Augraphy scanner noise to RVL-CDIP-N test images (ablation)VGG-16 · RVL-CDIP-Nmicro accuracy: 0.668 -> 0.650No clear effect
add Augraphy scanner noise to RVL-CDIP-N test images (ablation)ResNet-50 · RVL-CDIP-Nmicro accuracy: 0.598 -> 0.606No clear effect
switch in-domain set from RVL-CDIP test to RVL-CDIP-N for OOD detectionDiT-base · RVL-CDIP-O detectionAUC (MSP): 0.894 -> 0.728Hurt
switch in-domain set from RVL-CDIP test to RVL-CDIP-N for OOD detectionDiT-base · RVL-CDIP-O detectionAUC (Energy): 0.888 -> 0.753Hurt
switch in-domain set from RVL-CDIP test to RVL-CDIP-N for OOD detectionVGG-16 · RVL-CDIP-O detectionAUC (Energy): 0.923 -> 0.648Hurt
use Energy instead of MSP confidence scoringDiT-base · RVL-CDIP-N vs RVL-CDIP-OAUC: 0.728 -> 0.753Improved
use Energy instead of MSP confidence scoringResNet-50 · RVL-CDIP-N vs RVL-CDIP-OAUC: 0.581 -> 0.554Hurt

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