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

In this article an innovative method for training regressive MLP networks ispresented, which is not subject to local minima. The Error-Back-Propagationalgorithm, proposed by William-Hinton-Rummelhart, has had the merit offavouring the development of machine learning techniques, which has permeatedevery branch of research and technology since the mid-1980s. This extraordinarysuccess is largely due to the black-box approach, but this same factor was alsoseen as a limitation, as soon more challenging problems were approached. One ofthe most critical aspects of the training algorithms was that of local minimaof the loss function, typically the mean squared error of the output on thetraining set. In fact, as the most popular training algorithms are driven bythe derivatives of the loss function, there is no possibility to evaluate if areached minimum is local or global. The algorithm presented in this paperavoids the problem of local minima, as the training is based on the propertiesof the distribution of the training set, or better on its image internal to theneural network. The performance of the algorithm is shown for a well-knownbenchmark.

A free from local minima algorithm for training regressive MLP neural networks