What are "multilayer shallow neural networks"?

In several documentation pages, Mathworks mentions "multilayer shallow neural networks" (NN), but I cannot understand what they mean.

Namely, I think 99% of people would agree that a shallow NN is one with only one hidden layer, whereas a deep NN is one with >1 hidden layer. Then, what is a multilayer shallow network? it seems a nonsense term.

examples:

• https://www.mathworks.com/help/deeplearning/ug/multilayer-neural-networks-and-backpropagation-training.html
• https://www.mathworks.com/help/deeplearning/ug/multilayer-neural-network-architecture.html
• https://www.mathworks.com/help/deeplearning/ug/train-and-apply-multilayer-neural-networks.html

Answer 1

[Drew]

The short answer is that "multilayer shallow neural networks" refers to neural networks that have, for example, 2 or 3 layers.

For example, on the second documentation page that you referenced (https://www.mathworks.com/help/deeplearning/ug/multilayer-neural-network-architecture.html), the last neural network shown on that page is a "two-layer tansig/purelin network".

[Marco Coraggio]

I see, thank you!