This is a simple artificial neural network simulator. Knowledge of neural networks is not required to use it, and you may learn something just by playing with a few networks.

Click on a neuron to artificially toggle its state.

What is happening?

Neurons, represented by the circles, are connected to other neurons by synapses, each with a randomly generated weight value.

When a neuron is activated, all of its outward-facing synapses are activated as well. Each of the connected neurons respond by calculating the sum of the weights of its receiving synapses. If this sum is greater than 1, the neuron will automatically be activated, possibly causing a chain reaction or periodic activity. The weights of the synapses may be large enough to trigger a neuron on its own, while some require additional help from other synapses. Others may even subtract from the total weights of a neuron and possibly deactivate it.

Features for later