AI Paper Atlas

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

ZeroSiam: An Efficient Siamese for Test-Time Entropy Optimization without Collapse

Guohao Chen, Shuaicheng Niu, Deyu Chen, Jiahao Yang, Zitian Zhang, Mingkui Tan, Pengcheng Wu, Zhiqi Shen

ICLR 2026 · Label-efficient and robust learning › Test-time adaptation · approach: Realistic and stable test-time adaptation

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 predictor with stop-gradient asymmetric alignment to Tent (ablation)Tent · ImageNet-C level 5 online label shifts, ViT-Baseaccuracy (%): 47.3 -> 64.1Improved
use a fixed random predictor instead of a learnable identity-initialized predictor (ablation)ZeroSiam · ImageNet-C level 5 online label shifts, ViT-Baseaccuracy (%): 64.1 -> 60.7Hurt
make the predictor deeper and nonlinear (FC+ReLU+FC) (ablation)ZeroSiam · ImageNet-C level 5 online label shifts, ViT-Baseaccuracy (%): 64.1 -> 64.0No clear effect
add ZeroSiam to EATAEATA · ImageNet-C level 5 label shifts (ViT-B, R-50, ConvNeXt-T avg)accuracy (%): 40.6 -> 49.0Improved
add ZeroSiam to DeYODeYO · ImageNet-C level 5 label shifts (ViT-B, R-50, ConvNeXt-T avg)accuracy (%): 46.5 -> 53.9Improved
swap objective or divergence (KL, rKL, JS, MSE, sKL) in the alignment term (ablation)ZeroSiam · ImageNet-C level 5 label shifts, ViT-Base, Tent lossaccuracy (%): 63.8 to 64.1 across divergencesNo clear effect
add the asymmetric predictor to the SLR objective (ablation)SLR · ImageNet-C level 5 label shifts, ViT-Baseaccuracy (%): 44.0 -> 63.2Improved
set alpha to 1.0 versus larger values (ablation)ZeroSiam · ImageNet-C level 5 online label shifts, ViT-Baseaccuracy (%): 64.1 -> 62.4 (alpha 1.0 -> 2.5)Hurt
vary predictor learning rate between 0.001 and 0.01 (ablation)ZeroSiam · ImageNet-C level 5 label shifts, ViT-Baseaccuracy (%): about 1% changeNo clear effect
use ZeroSiam instead of DeYO on a weak model under label shiftsViT-Small on noise corruptions · ImageNet-C level 5 label shiftsaccuracy (%): 0.1 -> 26.5Improved
adapt only on misclassified blind-spot samples with ZeroSiam instead of prior methodsImageNet-C level 5 batch size 1 · blind-spot subsetaverage accuracy (%): 29.0 -> 52.0Improved
add online entropy-based TTA (Tent, SAR) to Llama3.1-8B reasoning on CollegeMathLlama3.1-8B-Instruct · CollegeMathaccuracy (%): 25.00 -> 24.17 (Tent), 25.00 -> 24.25 (SAR)Hurt

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