AI Paper Atlas

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

FOZO: Forward-Only Zeroth-Order Prompt Optimization for Test-Time Adaptation

Xingyu Wang, Tao Wang

CVPR 2026 · Label-efficient and robust learning › Test-time adaptation · approach: Pseudo-label test-time adaptation · 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
add deep-shallow feature statistics alignment loss (ablation)FOZO · ImageNet-C (5k, level 5) component ablationTop-1 accuracy (%): 57.3 -> 60.1Improved
add dynamic perturbation scale decay (ablation)FOZO · ImageNet-C (5k, level 5) component ablationTop-1 accuracy (%): 60.1 -> 62.7Improved
add zeroth-order prompt optimization with entropy loss (ablation)ViT-Base · ImageNet-C (5k, level 5) component ablationTop-1 accuracy (%): 55.1 -> 57.3Improved
dynamic perturbation speeds convergence (ablation)FOZO · ImageNet-C level 5 convergenceaccuracy vs timeImproved
FOZO vs FOA (CMA-ES prompts) at 2 forward passesforward-only TTA · ImageNet-C (5k, level 5, continual), 2 FPTop-1 accuracy (%): 58.13 -> 59.52Improved
FOZO vs ZOA at 2 forward passesforward-only TTA · ImageNet-C (5k, level 5, continual), 2 FPTop-1 accuracy (%): 58.56 -> 59.52Improved
FOZO vs backprop methods (EATA) at 28 forward passesTTA · ImageNet-C (5k, level 5, continual), 28 FP vs backpropTop-1 accuracy (%): 61.35 -> 62.67Improved
FOZO on quantized INT8 model vs FOAquantized ViT-Base · ImageNet-C (5k, level 5), INT8 quantized ViT-BaseTop-1 accuracy (%): 57.07 -> 58.00Improved
increase prompts from 3 to 8 (ablation)FOZO · ImageNet-C Gaussian noise level 5, hyperparametersTop-1 accuracy (%)No clear effect
increase batch size from 32 to 128 (ablation)FOZO · ImageNet-C Gaussian noise level 5, hyperparametersTop-1 accuracy (%): 62.3 -> 61.2Hurt
reduce batch size to 4 (ablation)FOZO · ImageNet-C Gaussian noise level 5, hyperparametersTop-1 accuracy (%): 62.3 -> 60.9Hurt

Later papers that compared with this method

1 later accepted paper reports FOZO as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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