Product Neuron

Activation is some variant of a product of incoming activations. They are used in Long Short Term Memory networks. Also see the discussion of Sigma Pi units in the original PDP Volumes.

Use Weights

If false (the default case), activation is a product of incoming activations.

If true, activation is a product of incoming activation / weight products, or (in the case of spiking neurons) post-synaptic-responses (see the spike responder docs).

Add Noise

If this is set to true, random values are added to the activation via a noise generator. The random values are added after the Product activation function is applied. For details how the noise generator works, click here.