English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Zero-cost proxies are nowadays frequently studied and used to search forneural architectures. They show an impressive ability to predict theperformance of architectures by making use of their untrained weights. Thesetechniques allow for immense search speed-ups. So far the joint search forwell-performing and robust architectures has received much less attention inthe field of NAS. Therefore, the main focus of zero-cost proxies is the cleanaccuracy of architectures, whereas the model robustness should play an evenlyimportant part. In this paper, we analyze the ability of common zero-costproxies to serve as performance predictors for robustness in the popularNAS-Bench-201 search space. We are interested in the single prediction task forrobustness and the joint multi-objective of clean and robust accuracy. Wefurther analyze the feature importance of the proxies and show that predictingthe robustness makes the prediction task from existing zero-cost proxies morechallenging. As a result, the joint consideration of several proxies becomesnecessary to predict a model's robustness while the clean accuracy can beregressed from a single such feature.

An Evaluation of Zero-Cost Proxies -- from Neural Architecture Performance to Model Robustness