Ramp Function

The ramp function is a function of a single real variable whose graph looks like a ramp: it outputs zero for every negative input and simply returns the input itself for every non-negative input, giving it a slope of one from the origin onward. In mathematics it is also known as the positive part of a number, and it appears under different names in different fields: in machine learning it is the ReLU, or rectified linear unit, activation function, drawing its name from an analogy with half-wave rectification in electrical engineering, and in statistics, when used as a likelihood function, it underlies the tobit model. Smoothed, differentiable versions of the ramp function also exist for uses where a sharp corner at zero is undesirable. This description is adapted from Wikipedia contributors under CC BY-SA 4.0; changes were made. https://creativecommons.org/licenses/by-sa/4.0/