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

Recent advances in artificial intelligence largely benefit from better neuralnetwork architectures. These architectures are a product of a costly process oftrial-and-error. To ease this process, we develop ArchExplorer, a visualanalysis method for understanding a neural architecture space and summarizingdesign principles. The key idea behind our method is to make the architecturespace explainable by exploiting structural distances between architectures. Weformulate the pairwise distance calculation as solving an all-pairs shortestpath problem. To improve efficiency, we decompose this problem into a set ofsingle-source shortest path problems. The time complexity is reduced fromO(kn^2N) to O(knN). Architectures are hierarchically clustered according to thedistances between them. A circle-packing-based architecture visualization hasbeen developed to convey both the global relationships between clusters andlocal neighborhoods of the architectures in each cluster. Two case studies anda post-analysis are presented to demonstrate the effectiveness of ArchExplorerin summarizing design principles and selecting better-performing architectures.

Visual Analysis of Neural Architecture Spaces for Summarizing Design Principles