Given the increasing number of applications of artificial intelligence (AI), the idea of using AI with music has been around for a long time and is much researched. Since music is a series of notes, it is a classic example of a time series dataset. Time series datasets have proven very useful recently in a number of forecast areas – stock markets, weather patterns, sales patterns, and other time-based datasets. Recurrent neural networks (RNNs) are one of the most models for working with time series datasets. A popular enhancement made to RNNs is called long short-term memory (LSTM) neurons. We'll be using LSTMs in this chapter to work with the music notes.
Multimedia processing, too, isn't a new topic. Earlier in this project series, we covered image processing in detail in multiple chapters. In this...