AI Paper Atlas

Research map / Architectures and efficiency / Neural architecture search

Data-Free Neural Architecture Search via Recursive Label Calibration

Zechun Liu, Zhiqiang Shen, Yun Long, Eric Xing, Kwang-Ting Cheng, Chas Leichner

ECCV 2022 · Architectures and efficiency › Neural architecture search · approach: Efficient NAS search strategies · 7 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 recursive label calibration and regional update to regularized synthesis (ablation)synthetic data · SPOS search space, ImageNetSpearman rank correlation: 0.87 -> 0.96Improved
synthesize data with regularization instead of random noise (ablation)synthetic data · DARTS space, CIFAR-10accuracy rank correlationImproved
add recursive label calibration to synthesized data (ablation)synthetic data · DARTS space, CIFAR-10accuracy rank correlationImproved
use label calibration when synthesizing data for DARTS search (ablation)DARTS · CIFAR-10top-1 error: 3.29 -> 2.68Improved
use label calibration for SPOS search data (ablation)SPOS · ImageNettop-1 accuracy: 74.2Improved
replace original ImageNet with synthesized data for DARTS archDARTS · CIFAR-10top-1 error: 2.76 -> 2.68No clear effect
replace original data with synthesized data for SPOSSPOS · ImageNettop-1 error: 25.7 -> 25.8No clear effect
replace original data with synthesized data for ProxylessNASProxylessNAS · ImageNettop-1 error: 24.4 -> 24.2Improved
transfer data-free DARTS architecture to ImageNetDARTS · ImageNettop-1 error: 26.7 -> 26.4Improved
use our synthesis instead of ADI for data-free pruningMetaPruning · ImageNettop-1 error: 39.3 -> 36.5Improved
use our synthesis instead of ADI for knowledge transferstudent ResNet-18 · CIFAR-10top-1 error: 6.74 -> 5.97Improved
use our synthesis instead of ADI for knowledge transfer at 250 epochsstudent ResNet-50 · ImageNettop-1 error: 26.2 -> 25.9Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 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