This blog documents the process of teaching an AI system to generate melodies from the Skiptune database. Expect experiments, mistakes, and occasional surprises, and at the end expect melodies that sound as if a human wrote them.
Problem to Be Solved The Skiptune database is a database of 83,000 melodies from around the world and across several centuries. These melodies are encoded in a special way (a pitch difference followed by a duration ratio), and we want to explore whether an AI model can learn melodic structure from this representation. The
Last week we got halfway through our exploration of the definition of notes and durations as “words” akin to words in language, and we finish up that discussion this week with the following figure: Let’s walk through this plot, which uses the exact same data as the one last week but plots the points in
This week we check to see if music is similar to human language in that it also follows Heaps’ Law. There is not a large body of work on music and Heaps’ Law, at least not compared to that of linguistics. But there is a small, coherent line of research in computational musicology that either
Today we’re going to learn about Heaps’ Law in linguistics and how it can help us determine if treating musical melodies as made up of “words” like human language is justified. While this topic is necessarily math-oriented and I need to use formulas to be precise, I’ll also add plain language explanations and use figures
People some ask me what struck us the most while doing the Skiptune project, and the answer has consistently been the strong impression that new patterns of two consecutive notes emerge once in while at an irregular rate. Wondering why, an hypothesis came to mind that, loosely stated, composers incorporate a new two-note pattern when