The Sigmoid function is a mathematical function having a characteristic "S"-shaped curve or sigmoid curve.
When the sigmoid function is used in AI Neural networks to calculate bias within the layers of a Neural network this function is repeated and calculated for each neuron. The function must be simplified (or squished) to result in a usable function that can be used to program an AI software program.
In order to simplify the use of these functions, sigmoid function weights and activations are organized into columns into vectors resulting in a matrix vector product which will be the bias vector, as a final step the sigmoid is wrapped around the vector function. The results of these vectors can be symbolically represented in a simplified function, which then can be used to communicate the full transition of activations from one layer of the neural network to the next in a Squishified Sigmoid function.
When the sigmoid function is used in AI Neural networks to calculate bias within the layers of a Neural network this function is repeated and calculated for each neuron. The function must be simplified (or squished) to result in a usable function that can be used to program an AI software program.
In order to simplify the use of these functions, sigmoid function weights and activations are organized into columns into vectors resulting in a matrix vector product which will be the bias vector, as a final step the sigmoid is wrapped around the vector function. The results of these vectors can be symbolically represented in a simplified function, which then can be used to communicate the full transition of activations from one layer of the neural network to the next in a Squishified Sigmoid function.
by Archii May 07, 2019