Hello,
I was looking to implement this model https://arxiv.org/pdf/2002.09571.pdf but as a spike neural network.
The paper details 2 neural networks, 1 neuromodulatory gating network, and 1 classifier. The output of the neuromodulatory gating network interacts with the classifier network via an element-wise operation prior to the final fully connected layer in the classifier.
I would imagine if I wanted to create an SNN analog, I would have to recreate the architecture of both neural networks. However, I'm not sure how I would implement the element-wise multiplication. Would I have to create that behavior in the spiking domain manually somewhere?