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

Echo State Network (ESN) presents a distinguished kind of recurrent neuralnetworks. It is built upon a sparse, random and large hidden infrastructurecalled reservoir. ESNs have succeeded in dealing with several non-linearproblems such as prediction, classification, etc. Thanks to its rich dynamics,ESN is used as an Autoencoder (AE) to extract features from original datarepresentations. ESN is not only used with its basic single layer form but alsowith the recently proposed Multi-Layer (ML) architecture. The well setting ofESN (basic and ML) architectures and training parameters is a crucial and hardlabor task. Generally, a number of parameters (hidden neurons, sparsity rates,input scaling) is manually altered to achieve minimum learning error. However,this randomly hand crafted task, on one hand, may not guarantee best trainingresults and on the other hand, it can raise the network's complexity. In thispaper, a hierarchical bi-level evolutionary optimization is proposed to dealwith these issues. The first level includes a multi-objective architectureoptimization providing maximum learning accuracy while sustaining thecomplexity at a reduced standard. Multi-objective Particle Swarm Optimization(MOPSO) is used to optimize ESN structure in a way to provide a trade-offbetween the network complexity decreasing and the accuracy increasing. Apareto-front of optimal solutions is generated by the end of the MOPSO process.These solutions present the set of candidates that succeeded in providing acompromise between different objectives (learning error and networkcomplexity). At the second level, each of the solutions already found undergo amono-objective weights optimization to enhance the obtained pareto-front.Empirical results ensure the effectiveness of the evolved ESN recurrent AEs(basic and ML) for noisy and noise free data.

Hierarchical Bi-level Multi-Objective Evolution of Single- and Multi-layer Echo State Network Autoencoders for Data Representations