description: a queue for the neurons that are ready to fire. the idea is to preform the computations for each layer as an input spike enters the network. for all neurons reached by the spike, update their membrane potential based on its current membrane potential and the timestamp of the last update. If it fires, add it to the "fire queue". This computation ripples through each layer. It assumes that there are few connections between layers.
to-do: figure out a structure for weights, membrane potentials and time-stamps. a connection is represented by a non-zero weight.
dod: this idea is implemented, tested and reviewed
sebastianmonten
commented
3 months ago
description: a queue for the neurons that are ready to fire. the idea is to preform the computations for each layer as an input spike enters the network. for all neurons reached by the spike, update their membrane potential based on its current membrane potential and the timestamp of the last update. If it fires, add it to the "fire queue". This computation ripples through each layer. It assumes that there are few connections between layers.
to-do: figure out a structure for weights, membrane potentials and time-stamps. a connection is represented by a non-zero weight.
dod: this idea is implemented, tested and reviewed