COALA: Co-Aligned Autoencoders for Learning Semantically Enriched Audio Representation

ICML 2020