AI Paper Atlas

Research map / Architectures and efficiency / Pruning and efficient tuning

Incremental Learning Through Deep Adaptation

Amir Rosenfeld, John K. Tsotsos

TPAMI 2018 · Architectures and efficiency › Pruning and efficient tuning · approach: Efficient ViT adaptation · 273 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 ImageNet-pretrained base network instead of Sketch-trained base (ablation)DAN · 8 datasetsmean top-1 accuracy: 83.7 -> 86.5Improved
use selected best base per dataset (ImageNet or Sketch) instead of single base (ablation)DAN · 8 datasetsmean top-1 accuracy: 86.46 -> 87.76Improved
use two base networks concatenated (Caltech-256 and Sketch) (ablation)DAN · 8 datasetsmean top-1 accuracy: same as DAN_sketchNo clear effect
use randomly initialized base network (ablation)DAN · 8 datasetsmean top-1 accuracy: 79.9 -> 76.3Hurt
use random base instead of Daimler-trained base (ablation)DAN · 8 datasetsmean top-1 accuracy: 75 -> 76.3Improved
use DAN controller from Sketch base instead of full fine-tuningDAN sketch · 8 datasetsmean top-1 accuracy: 84.2 -> 83.7No clear effect
fine-tune all layers instead of last layer only (ablation)Sketch base · 8 datasetsmean transfer accuracy: 50.3 -> 84.2Improved
freeze batch-norm during full fine-tuning (ft-full-bn-off) (ablation)Sketch base · 8 datasetsmean transfer accuracy: 84.2 -> 82.5Hurt
use DAN instead of Residual AdaptersDAN · Visual Decathlondecathlon score: 2118 -> 2851Improved
use DAN instead of Residual Adapters (joint training)DAN · Visual Decathlonmean accuracy: 77.17 -> 77.01Mixed
use DAN instead of vanilla fine-tuningDAN · Visual Decathlondecathlon score: 2500 -> 2851Improved
quantize DAN weights to 8 bits (ablation)DAN · 8 datasetsaccuracyNo 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