← Back to projects

Winner — MIT 6.S191 Intro to Deep Learning

Lab-winning RNN approach for next-note prediction in Irish folk music.

MIT • 2020

Python TensorFlow Deep Learning RNN
Winner — MIT 6.S191 Intro to Deep Learning cover image

I attended MIT’s Intro to Deep Learning course, which included several labs using TensorFlow.

One lab challenged students to develop and train an RNN to predict the next note in a corpus of Irish folk music.

My solution was unique in that it shuffled input songs to avoid learning note associations across song boundaries.

That approach won the lab competition.