Less is more: faster and better music version identification with embedding distillation

Serrà Julià, Joan; Yesiler, Furkan; Gómez Gutiérrez, Emilia, 1975-

Less is more: faster and better music version identification with embedding distillation

License

Citation

Yesiler F, Serrà J, Gómez E. Less is more: faster and better music version identification with embedding distillation. In: Cumming J, Ha Lee J, McFee B, Schedl M, Devaney J, McKay C, Zagerle E, de Reuse T, editors. Proceedings of the 21st International Society for Music Information Retrieval Conference; 2020 Oct 11-16; Montréal, Canada. [Canada]: ISMIR; 2020. p. 884-92.

Abstract

Version identification systems aim to detect different renditions of the same underlying musical composition (loosely called cover songs). By learning to encode entire recordings into plain vector embeddings, recent systems have made significant progress in bridging the gap between accuracy and scalability, which has been a key challenge for nearly two decades. In this work, we propose to further narrow this gap by employing a set of data distillation techniques that reduce the embedding dimensionality of a pre-trained state-of-the-art model. We compare a wide range of techniques and propose new ones, from classical dimensionality reduction to more sophisticated distillation schemes. With those, we obtain 99% smaller embeddings that, moreover, yield up to a 3% accuracy increase. Such small embeddings can have an important impact in retrieval time, up to the point of making a real-world system practical on a standalone laptop.

Description

Comunicació presentada a: International Society for Music Information Retrieval Conference celebrat de l'11 al 16 d'octubre de 2020 de manera virtual.