AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

TransCL: Transformer Makes Strong and Flexible Compressive Learning

Chong Mou, Jian Zhang

TPAMI 2022 · Architectures and efficiency › Attention and transformers · approach: Efficient and robust vision transformers · 40 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 single-stage compressive learning instead of reconstruct-then-infer (ablation)TransCL-32 · ImageNet-1k val (CS ratio 1%)top-1 accuracy: 55.57 -> 78.00Improved
use single-stage compressive learning instead of reconstruct-then-infer (ablation)TransCL-32 · ImageNet-1k val (CS ratio 10%)top-1 accuracy: 80.65 -> 81.82Improved
learn the binary sampling matrix instead of float-point (ablation)TransCL-32 · ImageNet-1k val (CS ratio 1%)top-1 accuracy: 78.00 -> 76.12Hurt
learn the binary sampling matrix instead of float-point (ablation)TransCL-32 · ImageNet-1k val (CS ratio 10%)top-1 accuracy: 81.82 -> 80.91No clear effect
learn the binary sampling matrix instead of float-point (ablation)TransCL-16 · PASCAL Context (CS ratio 1%)mIoU: 50.83 -> 49.80Hurt
learn the binary sampling matrix instead of float-point (ablation)TransCL-16 · PASCAL Context (CS ratio 10%)mIoU: 55.52 -> 55.47No clear effect
replace CS measurement with bicubic downsampling and re-train ViT-B-32 (ablation)ViT-B-32 · ImageNet-1k val (CS ratio 1%)top-1 accuracy: 78.00 -> 73.91Hurt
use 32x32 blocks instead of 16x16 (ablation)TransCL · ImageNet-1k val (CS ratio 1%)top-1 accuracy: 78.86 -> 78.00Mixed
use transformer backbone instead of ResNet-152 (ResCL) (ablation)TransCL · ImageNet-1k val (CS ratio 10%)top-1 accuracy: 75.22 -> 81.82Improved
add knowledge distillation data-efficient training (ImageNet-1K only) (ablation)TransCL-16-1 · ImageNet-1k val (CS ratio 1%)top-1 accuracy: 78.86 -> 78.58No clear effect
reduce CS ratio from 10% to 1% (ablation)TransCL-16 · ImageNet-1k valtop-1 accuracy: 83.86 -> 78.86Hurt
reduce CS ratio from 10% to 1% (ablation)TransCL-16 · ADE20KmIoU: 50.00 -> 46.57Hurt

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