Woke up without sleeping as much as I wanted. Fine.
Wrote yet another bit about HMMs for the last chapter of my Master's work. This time it was about the training process of a HMM and the Baum-Welch algorithm. It seems to make perfect sense, even if you haven't seen the proof that it is a special case of the EM algorithm. It also seems to be slow.
I typed what I wrote.
And sarge is being updated again. Hooray