Recurrent neural networks and flow control mechanisms

JARNE, CECILIA; RODRIGO LAJE

Rungstedgaard

Conferencia; The Brain Conferences 2019, Dynamics of the brain: temporal aspects of computation; 2019

FENS, Federation of European Neuroscience Societies

Recurrent neural network models have been used in different fields during the last 35 years to help understand the fundamental characteristics of a great variety of dynamical systems. The main goal of this work is to present a simple model and a framework to perform a set of tasks and discuss the obtained results using this new framework. The proposed model represents a brain area or population of neurons that perform a sensory response to a stimulus, typically cortex where both recurrent inhibition and excitation occur. We trained our minimal recurrent neural network model to perform a series of tasks relevant to understand information processing and the control of information flow.