The website was built with the following languages:
- HTML
- CSS
- JavaScript
I did not use any special frameworks for this task. To be able to visualize the Spiking Neural Network, I took the visualization from Github user "Sdalas" as an approach. To make the SNN interactive, I decided to use the sliders to change the number of neurons from 0 to 20 and to strengthen the synapses from 0 to 20. My wish was to add more sliders to further exploit the formula, however, unfortunately, I found it very difficult. This is the reason why I could not visualize the graph (plot). But instead I understood how the SNN works. I will explain this in the next chapter.
Here I explain how the SNN works with the LIF model.
The special feature of the SNN is that it is closer to the way the brain works than conventional KNNs. Like the brain, SNNs work asynchronously, i.e. a neuron becomes active (red) only when the sum of the impulses stored in it exceeds the threshold in the membrane potential. This means in reverse that a neuron does not become active if it has a weak synapse strength.
In order to be able to understand this event, I will show several examples in the following.
SNN (10 neurons) with a strength of 2:
Here, neuron number 2 has been activated. The neuron has only one synapse, but it is strong in terms of strength. Thus, the following neuron number 1 is activated. Neuron number 1 also has one synapse. But the following neuron number 0 is not activated. This means that the strength of the synapse is very weak and doesn't reach the threshold. The continuation of this network stops.
The case is different when, for example, neuron number 9 is activated. It has two strong synapses this time and the following neurons also have one strong synapse each. Here, several neurons exceed or reach the threshold.
SNN (10 neurons) with a strength of 20:
Here, neuron number 3 has been activated, which in turn has activated all the others. The synapse strength is so strong that all neurons reach the threshold and can thus activate each other.