Implemented a model of spiking neurons using a Graph Neural Network with tunable parameters. The model can be simulated over time to give training data for the inverse problem. The framework allows for both randomly drawn and custom initial connectivity. There is also support for adding external stimulation, either as a custom callable or using the RegularStimulation or SinStimulation classes. These classes have parameters that can be tuned to increase or decrease activity in the network.
Implemented a model of spiking neurons using a Graph Neural Network with tunable parameters. The model can be simulated over time to give training data for the inverse problem. The framework allows for both randomly drawn and custom initial connectivity. There is also support for adding external stimulation, either as a custom callable or using the RegularStimulation or SinStimulation classes. These classes have parameters that can be tuned to increase or decrease activity in the network.