Composing Bach Chorales Using Deep Learning

Explore the intersection of artificial intelligence and musical creativity in this GOTO Copenhagen 2019 conference talk. Dive into the world of BachBot, an AI system using deep learning and long short-term memory (LSTM) to compose music in the style of Johann Sebastian Bach. Learn about the largest musical Turing test conducted to date, which revealed that listeners could only distinguish BachBot's compositions from real Bach pieces 5% better than random guessing. Discover the potential of algorithmic composition in Bach chorales and gain insights into the application of machine learning in computational musicology. Examine topics such as music notation, polyphony, recurrent neural networks, feedforward neural networks, and the intricacies of training AI models for musical composition. Uncover fascinating research findings and explore the exciting possibilities at the intersection of technology and classical music.

Intro
Overview
Music Notation
Polyphony
Recurrent Neural Networks
Feedforward Neural Networks
Juggling Weights
Music
Research Findings
The Most Exciting Part