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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Convolutional Neural Networks (CNNs) continue to achieve great success inclassification tasks as innovative techniques and complex multi-patharchitecture topologies are introduced. Neural Architecture Search (NAS) aimsto automate the design of these complex architectures, reducing the need forcostly manual design work by human experts. Cellular Encoding (CE) is anevolutionary computation technique which excels in constructing novelmulti-path topologies of varying complexity and has recently been applied withNAS to evolve CNN architectures for various classification tasks. However,existing CE approaches have severe limitations. They are restricted to only onedomain, only partially implement the theme of CE, or only focus on themicro-architecture search space. This paper introduces a new CE representationand algorithm capable of evolving novel multi-path CNN architectures of varyingdepth, width, and complexity for image and text classification tasks. Thealgorithm explicitly focuses on the macro-architecture search space.Furthermore, by using a surrogate model approach, we show that the algorithmcan evolve a performant CNN architecture in less than one GPU day, therebyallowing a sufficient number of experiment runs to be conducted to achievescientific robustness. Experiment results show that the approach is highlycompetitive, defeating several state-of-the-art methods, and is generalisableto both the image and text domains.

XC-NAS: A New Cellular Encoding Approach for Neural Architecture Search of Multi-path Convolutional Neural Networks