In a graduate "Ecology & Evolution" class I am co-teaching this semester I, on a lark, asked the students to write a Haiku on the Expectation-Maximization (EM) algorithm.

I introduced it in class to talk about finding maximum-likelihood answers to some complex problems using a simple approach. Earlier I gave them some homework problems to find answers to by writing EM algorithms in R. We had a midterm exam coming up and there is not enough time in-class for the students to program a new algorithm in R for an exam question. Rather, I asked them to write a Haiku about it to see what they would write and if this gave me any insight into their understanding of the method. I like the results so much I am posting them here.

Data and a guess
Update guess with each cycle
Find the true value.

Hey, start with a guess
then plug it in again and again
and again and again, yay!

Just a few short steps
Almost silly how simple
We find the answer

Find Expectation:
Calculate the Maximum.
Re-evaluate.

Looping, elegant
Throw your guess into the loop
And see what comes out

We must use a guess
As the change becomes smaller
We approach the truth