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

Despite their successes in the field of self-learning AI, ConvolutionalNeural Networks (CNNs) suffer from having too many trainable parameters,impacting computational performance. Several approaches have been proposed toreduce the number of parameters in the visual domain, the Inceptionarchitecture [Szegedy et al., 2016] being a prominent example. This raises thequestion whether the number of trainable parameters in CNNs can also be reducedfor 1D inputs, such as time-series data, without incurring a substantial lossin classification performance. We propose and examine two methods forcomplexity reduction in AstroNet [Shallue & Vanderburg, 2018], a CNN forautomatic classification of time-varying brightness data of stars to detectexoplanets. The first method makes only a tactical reduction of layers inAstroNet while the second method also modifies the original input data by meansof a Gaussian pyramid. We conducted our experiments with various degrees ofdropout regularization. Our results show only a non-substantial loss inaccuracy compared to the original AstroNet, while reducing training time up to85 percent. These results show potential for similar reductions in other CNNapplications while largely retaining accuracy.

Reducing Artificial Neural Network Complexity: A Case Study on Exoplanet Detection