AI Paper Atlas

Research map / Architectures and efficiency / Neural architecture search

Pi-NAS: Improving Neural Architecture Search by Reducing Supernet Training Consistency Shift

Jiefeng Peng, Jiqi Zhang, Changlin Li, Guangrun Wang, Xiaodan Liang, Liang Lin

ICCV 2021 · Architectures and efficiency › Neural architecture search · approach: Efficient NAS search strategies · 18 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 cross-path learning and nontrivial mean teacher to SPOS (full Pi-NAS) (ablation)SPOS supernet · Ranking correlation (classification)Kendall's Tau: 0.19 -> 0.79Improved
add mean teacher to the supernet-Pi model (ablation)Pi-NAS · Ranking correlation (classification)Kendall's Tau: 0.48 -> 0.79Improved
add cross-path learning (ablation)Pi-NAS · Ranking correlation (classification)Kendall's Tau: 0.14 -> 0.79Improved
add downsample-sharing (ablation)Pi-NAS · Ranking correlation (classification)Kendall's Tau: 0.40 -> 0.79Improved
use nontrivial mean teacher with negative targets instead of trivial mean teacher (ablation)Pi-NAS · Supernet trainingKendall's Tau: collision -> 0.79Improved
replace SPOS/ResNeSt-50 with Pi-NAS-cls found architectureResNeSt-50 search space · ImageNet valtop-1 accuracy: 81.1 -> 81.6Improved
Pi-NAS vs SPOS supernet for architecture searchsearched models · ImageNet valtop-1 accuracy: 81.04 -> 81.6Improved
use Pi-NAS-trans backbone instead of ResNeSt-50Mask R-CNN · COCO 2017 valAP box: 42.81 -> 44.11Improved
use Pi-NAS-trans backbone instead of ResNeSt-50DeeplabV3 · ADE20K valmIoU: 45.18 -> 45.49Improved
use Pi-NAS-trans backbone instead of ResNeSt-50DeeplabV3 · Cityscapes valmIoU: 80.08 -> 80.40Improved
use Pi-NAS ranking for instance segmentation transfer vs SPOSsupernet · Ranking correlation (instance segmentation)Kendall's Tau: 0.18 -> 0.51Improved
apply Pi-NAS on NAS-Bench-201 instead of SPOSsupernet · NAS-Bench-201 CIFAR-10test accuracy: 93.57 -> 93.83Improved

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