AI Paper Atlas

Research map / Architectures and efficiency / Neural architecture search

Demystifying the Neural Tangent Kernel From a Practical Perspective: Can It Be Trusted for Neural Architecture Search Without Training?

Jisoo Mok, Byunggook Na, Ji-Hoon Kim, Dongyoon Han, Sungroh Yoon

CVPR 2022 · Architectures and efficiency › Neural architecture search · approach: Bi-level differentiable NAS · 18 citations (OpenAlex)

Check 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 LGA after one training epoch instead of the best-performing prior at-initialization NTK metricNAS architecture ranking · NAS-Bench-101Kendall's Tau: 0.094 -> 0.308Improved
Use LGA after one training epoch instead of the best-performing prior at-initialization NTK metricNAS architecture ranking · NDS-DARTSKendall's Tau: 0.234 -> 0.408Improved
Use LGA after one training epoch instead of the best-performing prior at-initialization NTK metricNAS architecture ranking · NDS-ENASKendall's Tau: 0.162 -> 0.416Improved
Use LGA after one training epoch instead of the best-performing prior at-initialization NTK metricNAS architecture ranking · NDS-NASNetKendall's Tau: 0.146 -> 0.357Improved
Use LGA after one training epoch instead of the best-performing prior at-initialization NTK metricNAS architecture ranking · NDS-AmoebaKendall's Tau: 0.158 -> 0.396Improved
Use LGA₅ rather than LGA₃Random search on NAS-Bench-201 · CIFAR-10test accuracy: 93.64 -> 94.03Improved
Use LGA₅ rather than LGA₃Random search on NAS-Bench-201 · CIFAR-100test accuracy: 69.77 -> 71.56Improved
Use LGA₅ rather than LGA₃Random search on NAS-Bench-201 · ImageNet-16-120test accuracy: 45.03 -> 46.30Improved
Use LGA₅ rather than LGA₃Evolutionary search on NAS-Bench-201 · CIFAR-10test accuracy: 94.30 -> 93.94Hurt
Use LGA₅ rather than LGA₃Evolutionary search on NAS-Bench-201 · CIFAR-100test accuracy: 71.18 -> 72.42Improved
Use LGA₅ rather than LGA₃Evolutionary search on NAS-Bench-201 · ImageNet-16-120test accuracy: 45.30 -> 45.17No clear effect
Use Gaussian initialization instead of Xavier or Kaiming initialization (ablation)F-Norm and Mean · NAS-Bench-201Kendall's TauHurt

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