Algorithms III: Markov Chains

In this module we look at a fairly common Machine Learning algorithm called the Markov Chain. We will look at it through the lens of creating musical compositions with the MIDI communication protocol. Music adds a new beguiling ingredient to Machine Learning problems: time. This consideration of time is a major component of many machine learning frameworks (LSTM, RNN) that deal with language, music, and even images. The relative simplicity of Markov Chains illustrate how dramatic the impact of “feeding back” or considering the past inputs of a model. So, after this module you should be able to:

  • Identify different types of Musical data
  • Understand the inter-workings and limitations of a Markov Chain
  • Illustrate the fundamental reason for and effects of incorporating previous inputs into a model
  • Determine the limitations and possibilities when working with different types of musical data