AI Paper Atlas

Research map / Architectures and efficiency / Neural architecture search

Adapting Neural Architectures Between Domains

Yanxi Li, Zhaohui Yang, Yunhe Wang, Chang Xu

NeurIPS 2020 · Architectures and efficiency › Neural architecture search · approach: Efficient NAS search strategies · 16 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 hybrid (target) loss in both training and search phases (ablation)AdaptNAS-S · ImageNet testtop-1 error (%): 25.3 -> 24.5Improved
add hybrid loss only in search phase (ablation)AdaptNAS-S · ImageNet testtop-1 error (%): 25.3 -> 24.8Improved
use cosine instead of exponential gamma scheduler (ablation)AdaptNAS-S · ImageNet testtop-1 error (%): 25.1 -> 25.3Hurt
use cosine instead of exponential gamma scheduler (ablation)AdaptNAS-S · CIFAR-10 testerror (%): 2.93 -> 2.77Improved
mix source and target loss (alpha 0.75, beta 0.98) instead of source loss only (ablation)AdaptNAS-C · ImageNet testtop-1 error (%): 25.7 -> 25.1Improved
target loss only (alpha 0) instead of hybrid at 1,000 target samples (beta 0.98) (ablation)AdaptNAS-C · ImageNet testtop-1 error (%): 25.7 -> 29.5Hurt
add target samples via AdaptNAS-C (alpha 0.25, beta 0.50) vs source-only (alpha 1.00) (ablation)AdaptNAS-C · ImageNet testtop-1 error (%): 25.4 -> 24.2Improved
add adversarial domain-distance loss during architecture search (AdaptNAS vs search on source)DARTS-style NAS · MNIST to SVHNtarget test accuracy (%): 94.70 -> 95.63Improved
add adversarial domain-distance loss (AdaptNAS vs search on source)DARTS-style NAS · MNIST to MNIST-Mtarget test accuracy (%): 98.56 -> 98.75Improved
use Rot-4 instead of Rot-1 rotation task for AdaptNAS-C (ablation)AdaptNAS-C · ImageNet testtop-1 error (%): 24.3 -> 24.2Improved
use Rot-4 instead of Rot-1 for AdaptNAS-S (ablation)AdaptNAS-S · ImageNet testtop-1 error (%): 24.7 -> 25.3Hurt

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