Hands-On Artificial Intelligence for Cybersecurity

CNNs

A CNN is a particular type of FFN, characterized by the fact that the organization of the neuronal layers follows the same organization of the existing visual apparatus in the biological world, with the superposition of regions of neurons within the visual field.

As we have said, in CNNs, each neuron is connected to a contiguous region of input neurons, in order to map the corresponding regions of pixels of an image.

In this way, it is possible to identify the spatial correlations through local connectivity schemes between neurons lying on adjacent layers, which allow, for example, the identification of objects.

In the CNNs, the contiguous regions of neurons are in fact organized by emulating the three-dimensional quantities of width, ...