Softmax activation function
- Softmax activation function
The softmax activation function is a neural transfer function. In neural networks, transfer functions calculate a layer's output from its net input. It is represented as:
Where "p" is the value of an output node, "q" is the net input to an output node, and "n" is the number of output nodes.
Wikimedia Foundation.
2010.
Look at other dictionaries:
Artificial neural network — An artificial neural network (ANN), usually called neural network (NN), is a mathematical model or computational model that is inspired by the structure and/or functional aspects of biological neural networks. A neural network consists of an… … Wikipedia
Backpropagation — Backpropagation, or propagation of error, is a common method of teaching artificial neural networks how to perform a given task. It was first described by Paul Werbos in 1974, but it wasn t until 1986, through the work of David E. Rumelhart,… … Wikipedia