A Continuous Spiking Neural Network

preview_player
Показать описание

A continuous spiking neural network is a spiking neural network where the neurons produce continuous outputs. The refractory period after a spike is emulated using a continuous variable called a discount, which is based on the output of a neuron. A discount is subtracted from the input of a neuron making it more difficult to activate if it has recently produced a high output. The discount decays over time until it reaches zero, at which point it no longer affects the neuron.
Рекомендации по теме
Комментарии
Автор

Very interesting work, I'm thrilled to see some code, thanks for sharing!

Mike
Автор

Hi Carson, this is really cool! Thanks for sharing the code. What library did you use for visualising the output (drawing nodes and edges)?

lewisb