AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

DREAM: Document Recognition with Explicit Adaptive Memory

Tianqi Zhao, Di Wu, Liangrui Peng, Yifan Huang, Kemeng Zhao, Shuo Li, Zhiyu Li, Yizhu Wang, Borui Jiang, Yuyang Li

CVPR 2026 · Architectures and efficiency › Attention and transformers · approach: Long-context KV and memory attention

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
add multiscale prototype memory with sparsity lossGOT-based document recognizer · Fox (English)Edit Distance: 0.107 -> 0.082Improved
add multiscale prototype memory with sparsity lossGOT-based document recognizer · Fox (Chinese)Edit Distance: 0.112 -> 0.101Improved
add DREAM memoryGOT-based document recognizer · DreamDoc testEdit Distance: 0.254 -> 0.248Improved
add DREAM memoryGOT-based document recognizer · DreamDoc testF1-score: 0.868 -> 0.869No clear effect
add DREAM memory to handwriting modelhandwriting text line recognizer · SCUT-HCCDocAR (%): 92.78 -> 93.36Improved
add DREAM memory to handwriting modelhandwriting text line recognizer · SCUT-EPTAR (%): 78.48 -> 78.82Improved
use multiscale memory instead of single-scale memory (ablation)DREAM · Fox (Chinese)Edit Distance: 0.126 -> 0.101Improved
add entropy sparsity loss to multiscale memory (ablation)DREAM · Fox (Chinese)Edit Distance: 0.104 -> 0.101Improved
add entropy sparsity loss to multiscale memory (ablation)DREAM · Fox (English)Edit Distance: 0.096 -> 0.082Improved
add global negative entropy term to local entropy loss (ablation)DREAM · Fox (Chinese)Edit Distance: 0.136 -> 0.101Improved
add global negative entropy term to local entropy loss (ablation)DREAM · Fox (English)Edit Distance: 0.087 -> 0.082Improved
use single-scale memory with sparsity loss instead of no memory (ablation)GOT-based document recognizer · Fox (Chinese)Edit Distance: 0.112 -> 0.126Hurt

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