Unsupervised learning of low-level audio features for music similarity estimation

Abstract

While there is an enormous amount of music data available, the eld of music analysis almost exclusively uses manually designed features. In this work we learn features from music data in a completely unsupervised way and evaluate them on a musical genre classi- cation task. We achieve results very close to state-of-the-art performance which relies on highly hand-tuned feature extractors.