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 Plus annual

Presents the access of premium content as premium feature

Premium Content

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 Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

In neural architecture search, the structure of the neural network to bestmodel a given dataset is determined by an automated search process. EfficientNeural Architecture Search (ENAS), proposed by Pham et al. (2018), has recentlyreceived considerable attention due to its ability to find excellentarchitectures within a comparably short search time. In this work, which ismotivated by the quest to further improve the learning speed of architecturesearch, we evaluate the learning progress of the controller which generates thearchitectures in ENAS. We measure the progress by comparing the architecturesgenerated by it at different controller training epochs, where architecturesare evaluated after having re-trained them from scratch. As a surprisingresult, we find that the learning curves are completely flat, i.e., there is noobservable progress of the controller in terms of the performance of itsgenerated architectures. This observation is consistent across the CIFAR-10 andCIFAR-100 datasets and two different search spaces. We conclude that the highquality of the models generated by ENAS is a result of the search space designrather than the controller training, and our results indicate that one-shotarchitecture design is an efficient alternative to architecture search by ENAS.

A Study of the Learning Progress in Neural Architecture Search Techniques