AI Paper Atlas

Research map / Architectures and efficiency / Neural architecture search

TRNAS: A Training-Free Robust Neural Architecture Search

Yeming Yang, Qingling Zhu, Jianping Luo, Ka-Chun Wong, Qiuzhen Lin, Jianqiang Li

ICCV 2025 · Architectures and efficiency › Neural architecture search · approach: Efficient NAS search strategies · 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
Use TRNAS searched architecture instead of ZCPRob under white-box attacksTRNAS · CIFAR-10AutoAttack accuracy (%): 49.97 -> 51.76Improved
Use TRNAS instead of ZCPRob on CIFAR-100TRNAS · CIFAR-100AutoAttack accuracy (%): 25.59 -> 27.32Improved
Transfer TRNAS architecture to ImageNetTRNAS · ImageNetAutoAttack accuracy (%): 8.81 -> 9.08 (vs SWAP)Improved
Use proxy-based search without any training instead of training-based robust NASTRNAS · CIFAR-10 search costGPU days: 0.02 vs 0.2 (CRoZe)Improved
Replace robust NTK proxy in ZCPRob baseline with R-Score (ablation)ZCPRob baseline · CIFAR-10 (Fig. 8)robust accuracyMixed
Add multi-objective selection (MOS) on top of R-Score (ablation)TRNAS · CIFAR-10 (Fig. 8)clean and robust accuracyImproved
Use R-Score instead of other zero-cost proxies (ablation)TRNAS · RobustBenchSpearman correlation with robustnessImproved
Use gradient-consistency proxy (CRoZe) compared to R-ScoreCRoZe · Proxy evaluation, 2000 architectureseffective rate: 79.30 % vs 100.0 %Hurt
Use classic clean-accuracy zero-cost proxies for robustnessZero-cost proxies · RobustBenchcorrelation with adversarial robustnessHurt
Use TRNAS searched architecture as black-box target and as sourceTRNAS · CIFAR-10 transfer attacksaccuracy under transfer attack: 65.57 target TRNAS vs 62.93 as sourceImproved

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