Enfoque de red neuronal de picos basado en el gusano Caenorhabditis Elegans para la clasificación
Fecha
Autor corporativo
Título de la revista
ISSN de la revista
Título del volumen
Editor
Compartir
Director
Altmetric
Resumen
A neural network is composed of a group of neurons that make a connectome, which is a map of neural connections that allows establishing the paths between the neurons. The neural network can generate the actions of living beings with the neuronal interaction through chemical-electrical signals. The behavior of machines is not dynamic compared to the behavior of animals; then, the machine’s behavior must be modeled and made by an exterior design, while in living beings, the behavior is caused by experience. Caenorhabditis Elegans is a worm model to study the connections of its neurons. In order to study the dynamic behavior in software systems based on biologic models, we created an approach to train and classify binary patterns using the structure of the Caenorhabditis Elegans’ connectome. We used the connectivity of neurons of Caenorhabditis Elegans to make a custom approach to train a Spiking Neural Network using a branching factor to classify patterns instead of layers of neurons. We made a software system to show the graph of neuronal connections of the Caenorhabditis Elegans. We also used Spike-Timing-Dependent Plasticity in order to establish the strength of the weights between the connections. In addition, we used a Hodgkin- Huxley model to calculate the neuron’s potential membrane and handle the spikes of the network.