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 evolutionary robotics an encoding of the control software, which mapssensor data (input) to motor control values (output), is shaped by stochasticoptimization methods to complete a predefined task. This approach is assumed tobe beneficial compared to standard methods of controller design in those caseswhere no a-priori model is available that could help to optimize performance.Also for robots that have to operate in unpredictable environments, anevolutionary robotics approach is favorable. We demonstrate here that such amodel-free approach is not a free lunch, as already simple tasks can representunsolvable barriers for fully open-ended uninformed evolutionary computationtechniques. We propose here the 'Wankelmut' task as an objective for anevolutionary approach that starts from scratch without pre-shaped controllersoftware or any other informed approach that would force the behavior to beevolved in a desired way. Our focal claim is that 'Wankelmut' represents thesimplest set of problems that makes plain-vanilla evolutionary computationfail. We demonstrate this by a series of simple standard evolutionaryapproaches using different fitness functions and standard artificial neuralnetworks as well as continuous-time recurrent neural networks. All our testedapproaches failed. We claim that any other evolutionary approach will also failthat does per-se not favor or enforce modularity and does not freeze or protectalready evolved functionalities. Thus we propose a hard-to-pass benchmark andmake a strong statement for self-complexifying and generative approaches inevolutionary computation. We anticipate that defining such a 'simplest task tofail' is a valuable benchmark for promoting future development in the field ofartificial intelligence, evolutionary robotics and artificial life.

Sooner than Expected: Hitting the Wall of Complexity in Evolution