
Feature Extraction and Machine Learning
Michael Scott Asato Cuthbert teaches this session of MIT's 21M.383 Computational Music Theory and Analysis, recorded in Spring 2023. The class opens with questions on the current problem set and discussion of final project expectations before turning to its main topic: feature extraction as a bridge between music analysis and machine learning. Cuthbert introduces how musical data, from pitch and rhythm to larger structural patterns, can be converted into features suitable for computational and statistical models, setting up how artificial intelligence techniques apply to music analysis. The seventy minute session mixes practical course logistics with conceptual grounding in computational methods, aimed at students who already have a background in both music theory and programming. It serves as a bridge lecture, connecting earlier analytical work in the course to the machine learning applications that follow in later classes.