majorado/IA

I invented a Mathematical Neuron model Y = Hz (Σ (weight * input) + bias) Activation of the layers in the frequencies of 8hz, 6hz and 4hz. Objective here is to enable "scalar learning" or SLM. With the reduction of the gradient, we obtain exponential transformation of the data, we can even make the matrix inversion diagonally and work with a sphere of the data volume. Deep Learning in exponential growth!