Expectation Maximization

So I spent the day the other day trying to get a hold of some good EM tutorials. Here's a list of some papers which I found useful: I found the presentation of EM in this tutorial by Sean Borman to be the most straightforward, even though the derivation which is described avoids some intuition and motivation behind the algorithm.

Hemant Tagare's tutorial, particularly section 3.1, does a good job of describing how we can obtain the likelihood of a distribution of observed data given a model which incorporates unknown data as well. And from this we derive the EM process.

Finally, Aaron D'Souza EM paper serves as a quick summary and cheat sheet of EM as applied to Mixtures of Gaussians.