AI Paper Atlas

Research map / Architectures and efficiency / Neural architecture search

Zen-NAS: A Zero-Shot NAS for High-Performance Image Recognition

Ming Lin, Pichao Wang, Zhenhong Sun, Hesen Chen, Xiuyu Sun, Qi Qian, Hao Li, Rong Jin

ICCV 2021 · Architectures and efficiency › Neural architecture search · approach: One-shot and few-shot supernets · 197 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 Zen-Score instead of FLOPs as NAS proxy (ablation)Zen-NAS evolutionary search · CIFAR-100top-1 accuracy: 64.7% -> 80.1%Improved
use Zen-Score instead of synflow as NAS proxy (ablation)Zen-NAS evolutionary search · CIFAR-100top-1 accuracy: 75.9% -> 80.1%Improved
use Zen-Score instead of synflow as NAS proxy (ablation)Zen-NAS evolutionary search · CIFAR-10top-1 accuracy: 95.1% -> 96.2%Improved
use Zen-Score instead of TE-Score as NAS proxy (ablation)Zen-NAS evolutionary search · CIFAR-10top-1 accuracy: 96.1% -> 96.2%Improved
use Zen-Score instead of NASWOT as NAS proxy (ablation)Zen-NAS evolutionary search · CIFAR-100top-1 accuracy: 77.5% -> 80.1%Improved
use FLOPs or gradient-norm as proxy instead of random search (ablation)NAS proxy · CIFAR-10top-1 accuracy: 93.5% -> 93.1%Hurt
use Zen-Score instead of TE-Score for scoring speedResNet-50 scoring · time for 16 images at 224x224time (s): 0.77 -> 0.037Improved
use Zen-Score instead of NASWOT for scoring speedResNet-18 scoring · time for 16 images at 224x224time (s): 0.040 -> 0.012Improved
search with Zen-NAS instead of EfficientNet scaling (accuracy vs cost)ImageNet ZenNet-1.2ms · ImageNet-1k, V100 FP16 batch 64top-1 accuracy / search cost: 83.6 at 0.5 GPU days vs 84.4 at 3800 GPU daysImproved
add enhanced training recipe to ResNet-152ResNet-152 · ImageNet-1k valtop-1 accuracy: 79.2% -> 82.3%Improved
add BN layers back to deep vanilla network without rescaling (ablation)Phi-score · networks of varying depth/widthscore discriminationHurt
add BN-variance rescaling (Zen-Score) to the BN-network score (ablation)Phi-score · deep vanilla networksnumerical stability: overflow after 30 layers without BNImproved

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