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

To increase the ubiquity of machine learning it needs to be automated.Automation is cost-effective as it allows experts to spend less time tuning theapproach, which leads to shorter development times. However, while thisautomation produces highly accurate architectures, they can be uninterpretable,acting as `black-boxes' which produce low conventional errors but fail to modelthe underlying input-output relationships -- the ground truth. This paperexplores the use of the Fit to Median Error measure in machine learningregression automation, using evolutionary computation in order to improve theapproximation of the ground truth. When used alongside conventional errormeasures it improves interpretability by regularising learnt input-outputrelationships to the conditional median. It is compared to traditionalregularisers to illustrate that the use of the Fit to Median Error producesregression neural networks which model more consistent input-outputrelationships. The problem considered is ship power prediction using afuel-saving air lubrication system, which is highly stochastic in nature. Thenetworks optimised for their Fit to Median Error are shown to approximate theground truth more consistently, without sacrificing conventional Minkowski-rerror values.

Automation for Interpretable Machine Learning Through a Comparison of Loss Functions to Regularisers