AI Paper Atlas

Research map / Label-efficient and robust learning / Test-time adaptation

Convolutional Visual Prompt for Robust Visual Perception

Yun-Yun Tsai, Chengzhi Mao, Junfeng Yang

NeurIPS 2023 · Label-efficient and robust learning › Test-time adaptation · approach: Test-time adaptation of vision-language models · 2 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
Replace unstructured visual prompting with a random-initialized 3×3 convolutional promptWideResNet18 · CIFAR-10-Caverage error (%): 58.24 -> 52.37; -5.87Improved
Replace unstructured visual prompting with a random-initialized 3×3 convolutional promptResNet26 · CIFAR-10-Caverage error (%): 62.12 -> 59.32; -2.80Improved
Apply CVP with a random-initialized 3×3 kernelResNet50 · ImageNet-CmCE: 76.87 -> 75.34; -1.49Improved
Apply CVP with a random-initialized 5×5 kernelResNet50 · ImageNet-Rerror (%): 63.83 -> 63.06; -0.77Improved
Apply CVP with a random-initialized 3×3 kernelResNet50 · ImageNet-Sketcherror (%): 75.90 -> 75.30; -0.60Improved
Apply CVP with a random-initialized 5×5 kernelResNet50 · ImageNet-Aerror (%): 100.0 -> 98.4; -1.6Improved
Use a convolutional prompt instead of a rank-3 low-rank visual prompt (ablation)Visual prompting · CIFAR-10-Caverage error (%): 54.86 -> 52.37; -3.38 vs. standard for LVP, -5.87 for CVPImproved
Use unstructured rather than convolutional prompting with supervised labels (ablation)Visual prompting · CIFAR-10-Caverage error (%): 49.59 -> 42.02Improved
Add CVP input adaptation on top of TENT weight adaptation (ablation)TENT · CIFAR-10-Caverage error (%): 38.52 -> 36.69; -1.83Improved
Add CVP input adaptation on top of TTT (ablation)TTT · CIFAR-10-Caverage error (%): 52.92 -> 53.07Hurt
Use MAE rather than contrastive learning as the self-supervised objective (ablation)CVP · ImageNet-C, severity 1robust accuracy: 58.97 -> 60.94Improved
Disable test-time kernel updates (ablation)CVP with random 3×3 kernel · CIFAR-10-Caverage error (%): 53.77 -> 52.37Hurt

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