Neural Networks vs. Deep Neural Networks?
Units, weighted connections, and activation functions. Listen, watch and propagate millions of pixels and soundwaves into algorithms.
What’s makes a Neural Network?
Neural networks (NN) are always made up of these things called “units”, “weighted connections” which are between units and “activation functions”, which act just like neurons except with values for each individual unit itself and a degree of specific activation systematic properties which are determined by the propagated values along the network.
Each unit is excited or activated by the building and training from deep learning; pixel colors, sound pitch, etc.
Years ago, “deep networks” could be 3-5 layers; modern deep networks have become more popular and not to long ago around 2015 deep networks could consist of millions of “Input units”, to differentiate colors, voice, or the object in view to compare to the “hidden units/layers,” characterized aspects, assets, and “output units” - which would then clarify the classe(s): person, animal or instrument.
