logo
ResearchBunny Logo
Abstract
This study investigates which neural network architecture (feedforward vs. recurrent) best matches human behavior in artificial grammar learning. Human subjects and both feedforward and recurrent networks were tested on four grammars of varying complexity. Results show that both architectures can learn the grammars, but recurrent networks perform closer to human behavior, particularly with simpler, more explicit grammars, suggesting recurrent networks better model explicit learning while feedforward networks may capture implicit learning dynamics.
Publisher
Scientific Reports
Published On
Dec 17, 2020
Authors
Andrea Alamia, Victor Gauducheau, Dimitri Paisios, Rufin VanRullen
Tags
artificial grammar learning
neural networks
recurrent architecture
feedforward architecture
human behavior

Related Publications

Explore these studies to deepen your understanding of the subject.

Interpretable early warning recommendations in interactive learning environments: a deep-neural network approach based on learning behavior knowledge graph
Education

Interpretable early warning recommendations in interactive learning environments: a deep-neural network approach based on learning behavior knowledge graph

X. Xia and W. Qi

Neural and computational underpinnings of biased confidence in human reinforcement learning
Psychology

Neural and computational underpinnings of biased confidence in human reinforcement learning

C. Ting, N. Salem-garcia, et al.

Exercise intervention in middle-aged and elderly individuals with insomnia improves sleep and restores connectivity in the motor network
Health and Fitness

Exercise intervention in middle-aged and elderly individuals with insomnia improves sleep and restores connectivity in the motor network

R. Chen, S. Wang, et al.

In-sensor human gait analysis with machine learning in a wearable microfabricated accelerometer
Engineering and Technology

In-sensor human gait analysis with machine learning in a wearable microfabricated accelerometer

G. Dion, A. Tessier-poirier, et al.

Listen, Learn & Level Up
Over 10,000 hours of research content in 25+ fields, available in 12+ languages.
No more digging through PDFs—just hit play and absorb the world's latest research in your language, on your time.
listen to research audio papers with researchbunny