Gradient ascent is an optimization algorithm used to find the maximum of a function by iteratively moving in the direction of the steepest ascent as defined by the gradient. This method is particularly useful in machine learning for maximizing likelihood functions or other objective functions that need to be maximized rather than minimized.