AI Paper Atlas

Research map / Continual learning / Online and replay-based

Attention Retention for Continual Learning with Vision Transformers

Yue Lu, Xiangyu Zhou, Shizhou Zhang, Yinghui Xing, Guoqiang Liang, Wencong Zhang

AAAI 2026 · Continual learning › Online and replay-based · approach: Theoretical analysis of forgetting · 0 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
add adaptive thresholding instead of fixed top-20% masks (raw attention) (ablation)ARCL-ViT · 10S-ImageNet-Raccuracy: 77.63 -> 80.33Improved
add adaptive thresholding instead of fixed top-20% masks (layer-wise rollout) (ablation)ARCL-ViT · 10S-ImageNet-Raccuracy: 76.72 -> 81.17Improved
add adaptive thresholding instead of fixed (layer-wise rollout) (ablation)ARCL-ViT · 10S-DomainNetforgetting: 21.61 -> 7.04Improved
use layer-wise rollout instead of raw attention (ablation)ARCL-ViT · 10S-ImageNet-Raccuracy: 80.33 -> 81.17Improved
use layer-wise rollout instead of naive rollout (ablation)ARCL-ViT · 10S-DomainNetaccuracy: 82.28 -> 83.94Improved
mask previous attention regions instead of random positions (ablation)ARCL-ViT · 10S-ImageNet-Raccuracy: 73.27 -> 81.17Improved
mask previous attention regions instead of non-attention regions (ablation)ARCL-ViT · 10S-DomainNetaccuracy: 76.08 -> 83.94Improved
mask random regions instead of attention regions (ablation)ARCL-ViT · 10S-ImageNet-Rforgetting: 4.22 vs 4.84No clear effect
apply gradient masking of attention regions to sequential fine-tuningSeq-FT · 10S-ImageNet-Raccuracy: 49.43 -> 81.17Improved
use ARCL-ViT instead of VPT-CPG with DINO-1k weightsARCL-ViT · 10S-ImageNet-R (DINO-1k)accuracy: 68.32 -> 72.37Improved
use ARCL-ViT instead of VPT-CPG with iBOT-1k weightsARCL-ViT · 10S-DomainNet (iBOT-1k)accuracy: 77.79 -> 80.88Improved
use ARCL-ViT instead of VPT-CPG on 100-task ImageNet-RARCL-ViT · 100S-ImageNet-Raccuracy: 64.63 -> 70.72Improved

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