It isn't often that you get the chance to view a video that includes a discussion of a wide range of very technical topics. In this interview, Brian Beckman really does range over topics as diverse as the Viterbi algorithm and the Higgs boson.
Brian Beckman is an ex-cosmologist who now works for Microsoft Research. He seems to have the luck to work on interesting things:
"His team works on innovative incubation projects that turn into developer platform features (like LINQ, for example) and more."
In this video he discusses the implementation of the Viterbi algorithm in C#. If you have no idea what the Viterbi algorithm is then it would take some time to explain it in detail but if you joined in with the Stanford online AI course you will probably know that it is all about Hidden Markov Models - HMM.
In a nutshell an HMM postulates a multi-state system for the production of some outcome with the probabilities of the outcomes varying according to the state. The Viterbi algorithm is a way of estimating the probabilities of the states given the observations. This is useful in a range of AI applications and in error correcting codes.
The video is an hour long, but it is worth it:
On the way, the topics covered include Rx, monads and the promised Higgs Boson.
If you want to download the code demos these are at:
Have you had a suspicion that your GPS app is overestimating the distance traveled? If so you are probably correct but the reason isn't an algorithmic glitch. The answer lies in the statistics and it [ ... ]