A phenomena where a perceptron algorithm effectively stabilizes, ensuring that it can find a solution for linearly separable datasets after a finite number of iterations.
Generality: 500