Automated curriculum learning for embodied agents a neuroevolutionary approach

Abstract We demonstrate how the evolutionary training of embodied agents can be extended with a curriculum learning algorithm that automatically selects the environmental conditions in which the evolving agents are evaluated. The environmental conditions are selected to adjust the level of difficult...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Nicola Milano, Stefano Nolfi
Formato: article
Lenguaje:EN
Publicado: Nature Portfolio 2021
Materias:
R
Q
Acceso en línea:https://doaj.org/article/526589e900f34ccc951da87100eb776a
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
id oai:doaj.org-article:526589e900f34ccc951da87100eb776a
record_format dspace
spelling oai:doaj.org-article:526589e900f34ccc951da87100eb776a2021-12-02T17:15:17ZAutomated curriculum learning for embodied agents a neuroevolutionary approach10.1038/s41598-021-88464-52045-2322https://doaj.org/article/526589e900f34ccc951da87100eb776a2021-04-01T00:00:00Zhttps://doi.org/10.1038/s41598-021-88464-5https://doaj.org/toc/2045-2322Abstract We demonstrate how the evolutionary training of embodied agents can be extended with a curriculum learning algorithm that automatically selects the environmental conditions in which the evolving agents are evaluated. The environmental conditions are selected to adjust the level of difficulty to the ability level of the current evolving agents, and to challenge the weaknesses of the evolving agents. The method does not require domain knowledge and does not introduce additional hyperparameters. The results collected on two benchmark problems, that require to solve a task in significantly varying environmental conditions, demonstrate that the method proposed outperforms conventional learning methods and generates solutions which are robust to variations and able to cope with different environmental conditions.Nicola MilanoStefano NolfiNature PortfolioarticleMedicineRScienceQENScientific Reports, Vol 11, Iss 1, Pp 1-14 (2021)
institution DOAJ
collection DOAJ
language EN
topic Medicine
R
Science
Q
spellingShingle Medicine
R
Science
Q
Nicola Milano
Stefano Nolfi
Automated curriculum learning for embodied agents a neuroevolutionary approach
description Abstract We demonstrate how the evolutionary training of embodied agents can be extended with a curriculum learning algorithm that automatically selects the environmental conditions in which the evolving agents are evaluated. The environmental conditions are selected to adjust the level of difficulty to the ability level of the current evolving agents, and to challenge the weaknesses of the evolving agents. The method does not require domain knowledge and does not introduce additional hyperparameters. The results collected on two benchmark problems, that require to solve a task in significantly varying environmental conditions, demonstrate that the method proposed outperforms conventional learning methods and generates solutions which are robust to variations and able to cope with different environmental conditions.
format article
author Nicola Milano
Stefano Nolfi
author_facet Nicola Milano
Stefano Nolfi
author_sort Nicola Milano
title Automated curriculum learning for embodied agents a neuroevolutionary approach
title_short Automated curriculum learning for embodied agents a neuroevolutionary approach
title_full Automated curriculum learning for embodied agents a neuroevolutionary approach
title_fullStr Automated curriculum learning for embodied agents a neuroevolutionary approach
title_full_unstemmed Automated curriculum learning for embodied agents a neuroevolutionary approach
title_sort automated curriculum learning for embodied agents a neuroevolutionary approach
publisher Nature Portfolio
publishDate 2021
url https://doaj.org/article/526589e900f34ccc951da87100eb776a
work_keys_str_mv AT nicolamilano automatedcurriculumlearningforembodiedagentsaneuroevolutionaryapproach
AT stefanonolfi automatedcurriculumlearningforembodiedagentsaneuroevolutionaryapproach
_version_ 1718381260135464960