Comparing feedforward and recurrent neural network architectures with human behavior in artificial grammar learning

Abstract In recent years artificial neural networks achieved performance close to or better than humans in several domains: tasks that were previously human prerogatives, such as language processing, have witnessed remarkable improvements in state of the art models. One advantage of this technologic...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Andrea Alamia, Victor Gauducheau, Dimitri Paisios, Rufin VanRullen
Formato:	article
Lenguaje:	EN
Publicado:	Nature Portfolio 2020
Materias:	Medicine R Science Q
Acceso en línea:	https://doaj.org/article/c2575fa6d2e64527acdf99f9febf6d42
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:c2575fa6d2e64527acdf99f9febf6d42
record_format	dspace
spelling	oai:doaj.org-article:c2575fa6d2e64527acdf99f9febf6d422021-12-02T12:42:18ZComparing feedforward and recurrent neural network architectures with human behavior in artificial grammar learning10.1038/s41598-020-79127-y2045-2322https://doaj.org/article/c2575fa6d2e64527acdf99f9febf6d422020-12-01T00:00:00Zhttps://doi.org/10.1038/s41598-020-79127-yhttps://doaj.org/toc/2045-2322Abstract In recent years artificial neural networks achieved performance close to or better than humans in several domains: tasks that were previously human prerogatives, such as language processing, have witnessed remarkable improvements in state of the art models. One advantage of this technological boost is to facilitate comparison between different neural networks and human performance, in order to deepen our understanding of human cognition. Here, we investigate which neural network architecture (feedforward vs. recurrent) matches human behavior in artificial grammar learning, a crucial aspect of language acquisition. Prior experimental studies proved that artificial grammars can be learnt by human subjects after little exposure and often without explicit knowledge of the underlying rules. We tested four grammars with different complexity levels both in humans and in feedforward and recurrent networks. Our results show that both architectures can “learn” (via error back-propagation) the grammars after the same number of training sequences as humans do, but recurrent networks perform closer to humans than feedforward ones, irrespective of the grammar complexity level. Moreover, similar to visual processing, in which feedforward and recurrent architectures have been related to unconscious and conscious processes, the difference in performance between architectures over ten regular grammars shows that simpler and more explicit grammars are better learnt by recurrent architectures, supporting the hypothesis that explicit learning is best modeled by recurrent networks, whereas feedforward networks supposedly capture the dynamics involved in implicit learning.Andrea AlamiaVictor GauducheauDimitri PaisiosRufin VanRullenNature PortfolioarticleMedicineRScienceQENScientific Reports, Vol 10, Iss 1, Pp 1-15 (2020)
institution	DOAJ
collection	DOAJ
language	EN
topic	Medicine R Science Q
spellingShingle	Medicine R Science Q Andrea Alamia Victor Gauducheau Dimitri Paisios Rufin VanRullen Comparing feedforward and recurrent neural network architectures with human behavior in artificial grammar learning
description	Abstract In recent years artificial neural networks achieved performance close to or better than humans in several domains: tasks that were previously human prerogatives, such as language processing, have witnessed remarkable improvements in state of the art models. One advantage of this technological boost is to facilitate comparison between different neural networks and human performance, in order to deepen our understanding of human cognition. Here, we investigate which neural network architecture (feedforward vs. recurrent) matches human behavior in artificial grammar learning, a crucial aspect of language acquisition. Prior experimental studies proved that artificial grammars can be learnt by human subjects after little exposure and often without explicit knowledge of the underlying rules. We tested four grammars with different complexity levels both in humans and in feedforward and recurrent networks. Our results show that both architectures can “learn” (via error back-propagation) the grammars after the same number of training sequences as humans do, but recurrent networks perform closer to humans than feedforward ones, irrespective of the grammar complexity level. Moreover, similar to visual processing, in which feedforward and recurrent architectures have been related to unconscious and conscious processes, the difference in performance between architectures over ten regular grammars shows that simpler and more explicit grammars are better learnt by recurrent architectures, supporting the hypothesis that explicit learning is best modeled by recurrent networks, whereas feedforward networks supposedly capture the dynamics involved in implicit learning.
format	article
author	Andrea Alamia Victor Gauducheau Dimitri Paisios Rufin VanRullen
author_facet	Andrea Alamia Victor Gauducheau Dimitri Paisios Rufin VanRullen
author_sort	Andrea Alamia
title	Comparing feedforward and recurrent neural network architectures with human behavior in artificial grammar learning
title_short	Comparing feedforward and recurrent neural network architectures with human behavior in artificial grammar learning
title_full	Comparing feedforward and recurrent neural network architectures with human behavior in artificial grammar learning
title_fullStr	Comparing feedforward and recurrent neural network architectures with human behavior in artificial grammar learning
title_full_unstemmed	Comparing feedforward and recurrent neural network architectures with human behavior in artificial grammar learning
title_sort	comparing feedforward and recurrent neural network architectures with human behavior in artificial grammar learning
publisher	Nature Portfolio
publishDate	2020
url	https://doaj.org/article/c2575fa6d2e64527acdf99f9febf6d42
work_keys_str_mv	AT andreaalamia comparingfeedforwardandrecurrentneuralnetworkarchitectureswithhumanbehaviorinartificialgrammarlearning AT victorgauducheau comparingfeedforwardandrecurrentneuralnetworkarchitectureswithhumanbehaviorinartificialgrammarlearning AT dimitripaisios comparingfeedforwardandrecurrentneuralnetworkarchitectureswithhumanbehaviorinartificialgrammarlearning AT rufinvanrullen comparingfeedforwardandrecurrentneuralnetworkarchitectureswithhumanbehaviorinartificialgrammarlearning
_version_	1718393708494192640

Comparing feedforward and recurrent neural network architectures with human behavior in artificial grammar learning

Ejemplares similares