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

The ongoing advancements in network architecture design have led toremarkable achievements in deep learning across various challenging computervision tasks. Meanwhile, the development of neural architecture search (NAS)has provided promising approaches to automating the design of networkarchitectures for lower prediction error. Recently, the emerging applicationscenarios of deep learning have raised higher demands for network architecturesconsidering multiple design criteria: number of parameters/floating-pointoperations, and inference latency, among others. From an optimization point ofview, the NAS tasks involving multiple design criteria are intrinsicallymultiobjective optimization problems; hence, it is reasonable to adoptevolutionary multiobjective optimization (EMO) algorithms for tackling them.Nonetheless, there is still a clear gap confining the related research alongthis pathway: on the one hand, there is a lack of a general problem formulationof NAS tasks from an optimization point of view; on the other hand, there arechallenges in conducting benchmark assessments of EMO algorithms on NAS tasks.To bridge the gap: (i) we formulate NAS tasks into general multi-objectiveoptimization problems and analyze the complex characteristics from anoptimization point of view; (ii) we present an end-to-end pipeline, dubbed$\texttt{EvoXBench}$, to generate benchmark test problems for EMO algorithms torun efficiently -- without the requirement of GPUs or Pytorch/Tensorflow; (iii)we instantiate two test suites comprehensively covering two datasets, sevensearch spaces, and three hardware devices, involving up to eight objectives.Based on the above, we validate the proposed test suites using sixrepresentative EMO algorithms and provide some empirical analyses. The code of$\texttt{EvoXBench}$ is available from$\href{https://github.com/EMI-Group/EvoXBench}{\rm{here}}$.

Neural Architecture Search as Multiobjective Optimization Benchmarks: Problem Formulation and Performance Assessment