# Discriminative vs Generative Methods that learn $p(y \mid x)$ directly are referred to as **discrimative** methods. They are referred to as discriminative because their goal is only to discriminate between different $ys. Methods that learn $p(x, y)$ are referred to as **generative** methods. They are called this because such methods could in principle *generate* $xs. --- Date: 20211124 Links to: [Machine-Learning](Machine-Learning.md) Tags: References: * []()