# Why Gaussian mixture model uses Expectation maximization instead of Gradient descent?

Why Gaussian mixture model uses Expectation maximization instead of Gradient descent?

What other models uses Expectation maximization to find best optimal parameters instead of using gradient descent?