Transfer Learning with Jukebox for Music Source Separation

Wadhah Zai El Amri; Oliver Tautz; Helge Ritter; Andrew Melnik

doi:10.1007/978-3-031-08337-2_35

Transfer Learning with Jukebox for Music Source Separation

Wadhah Zai El Amri^*, Oliver Tautz, Helge Ritter, Andrew Melnik

^*Corresponding author for this work

Bielefeld University

Research output: Chapter in book/report/conference proceeding › Conference contribution › Research › peer review

Abstract

In this work, we demonstrate how a publicly available, pre-trained Jukebox model can be adapted for the problem of audio source separation from a single mixed audio channel. Our neural network architecture, which is using transfer learning, is quick to train and the results demonstrate performance comparable to other state-of-the-art approaches that require a lot more compute resources, training data, and time. We provide an open-source code implementation of our architecture (https://github.com/wzaielamri/unmix ).

Original language	English
Title of host publication	Artificial Intelligence Applications and Innovations
Subtitle of host publication	18th IFIP WG 12.5 International Conference, AIAI 2022, Hersonissos, Crete, Greece, June 17–20, 2022, Proceedings, Part II
Editors	Ilias Maglogiannis, Lazaros Iliadis, John Macintyre, Paulo Cortez
Place of Publication	Cham
Publisher	Springer
Pages	426-433
Number of pages	8
Edition	1.
ISBN (Electronic)	978-3-031-08337-2
ISBN (Print)	978-3-031-08336-5, 978-3-031-08339-6
DOIs	https://doi.org/10.1007/978-3-031-08337-2_35
Publication status	Published - 10 Jun 2022
Externally published	Yes

Publication series

Name	IFIP Advances in Information and Communication Technology
Volume	647
ISSN (Print)	1868-4238
ISSN (Electronic)	1868-422X

Access to Document

10.1007/978-3-031-08337-2_35

Cite this

Zai El Amri, W., Tautz, O., Ritter, H., & Melnik, A. (2022). Transfer Learning with Jukebox for Music Source Separation. In I. Maglogiannis, L. Iliadis, J. Macintyre, & P. Cortez (Eds.), Artificial Intelligence Applications and Innovations : 18th IFIP WG 12.5 International Conference, AIAI 2022, Hersonissos, Crete, Greece, June 17–20, 2022, Proceedings, Part II (1. ed., pp. 426-433). (IFIP Advances in Information and Communication Technology; Vol. 647). Springer. https://doi.org/10.1007/978-3-031-08337-2_35

Zai El Amri, Wadhah ; Tautz, Oliver ; Ritter, Helge et al. / Transfer Learning with Jukebox for Music Source Separation. Artificial Intelligence Applications and Innovations : 18th IFIP WG 12.5 International Conference, AIAI 2022, Hersonissos, Crete, Greece, June 17–20, 2022, Proceedings, Part II. editor / Ilias Maglogiannis ; Lazaros Iliadis ; John Macintyre ; Paulo Cortez. 1. ed. Cham : Springer, 2022. pp. 426-433 (IFIP Advances in Information and Communication Technology).

@inproceedings{4b19c9abf8da452cbc73b39e0f32424a,

title = "Transfer Learning with Jukebox for Music Source Separation",

abstract = "In this work, we demonstrate how a publicly available, pre-trained Jukebox model can be adapted for the problem of audio source separation from a single mixed audio channel. Our neural network architecture, which is using transfer learning, is quick to train and the results demonstrate performance comparable to other state-of-the-art approaches that require a lot more compute resources, training data, and time. We provide an open-source code implementation of our architecture (https://github.com/wzaielamri/unmix ).",

author = "{Zai El Amri}, Wadhah and Oliver Tautz and Helge Ritter and Andrew Melnik",

note = "Publisher Copyright: {\textcopyright} 2022, IFIP International Federation for Information Processing.",

year = "2022",

month = jun,

day = "10",

doi = "10.1007/978-3-031-08337-2_35",

language = "English",

isbn = "978-3-031-08336-5",

series = "IFIP Advances in Information and Communication Technology",

publisher = "Springer",

pages = "426--433",

editor = "Ilias Maglogiannis and Lazaros Iliadis and John Macintyre and Paulo Cortez",

booktitle = "Artificial Intelligence Applications and Innovations",

edition = "1.",

}

Zai El Amri, W, Tautz, O, Ritter, H & Melnik, A 2022, Transfer Learning with Jukebox for Music Source Separation. in I Maglogiannis, L Iliadis, J Macintyre & P Cortez (eds), Artificial Intelligence Applications and Innovations : 18th IFIP WG 12.5 International Conference, AIAI 2022, Hersonissos, Crete, Greece, June 17–20, 2022, Proceedings, Part II. 1. edn, IFIP Advances in Information and Communication Technology, vol. 647, Springer, Cham, pp. 426-433. https://doi.org/10.1007/978-3-031-08337-2_35

Transfer Learning with Jukebox for Music Source Separation. / Zai El Amri, Wadhah; Tautz, Oliver; Ritter, Helge et al.
Artificial Intelligence Applications and Innovations : 18th IFIP WG 12.5 International Conference, AIAI 2022, Hersonissos, Crete, Greece, June 17–20, 2022, Proceedings, Part II. ed. / Ilias Maglogiannis; Lazaros Iliadis; John Macintyre; Paulo Cortez. 1. ed. Cham: Springer, 2022. p. 426-433 (IFIP Advances in Information and Communication Technology; Vol. 647).

Research output: Chapter in book/report/conference proceeding › Conference contribution › Research › peer review

TY - GEN

T1 - Transfer Learning with Jukebox for Music Source Separation

AU - Zai El Amri, Wadhah

AU - Tautz, Oliver

AU - Ritter, Helge

AU - Melnik, Andrew

PY - 2022/6/10

Y1 - 2022/6/10

N2 - In this work, we demonstrate how a publicly available, pre-trained Jukebox model can be adapted for the problem of audio source separation from a single mixed audio channel. Our neural network architecture, which is using transfer learning, is quick to train and the results demonstrate performance comparable to other state-of-the-art approaches that require a lot more compute resources, training data, and time. We provide an open-source code implementation of our architecture (https://github.com/wzaielamri/unmix ).

AB - In this work, we demonstrate how a publicly available, pre-trained Jukebox model can be adapted for the problem of audio source separation from a single mixed audio channel. Our neural network architecture, which is using transfer learning, is quick to train and the results demonstrate performance comparable to other state-of-the-art approaches that require a lot more compute resources, training data, and time. We provide an open-source code implementation of our architecture (https://github.com/wzaielamri/unmix ).

UR - http://www.scopus.com/inward/record.url?scp=85133305158&partnerID=8YFLogxK

U2 - 10.1007/978-3-031-08337-2_35

DO - 10.1007/978-3-031-08337-2_35

M3 - Conference contribution

SN - 978-3-031-08336-5

SN - 978-3-031-08339-6

T3 - IFIP Advances in Information and Communication Technology

SP - 426

EP - 433

BT - Artificial Intelligence Applications and Innovations

A2 - Maglogiannis, Ilias

A2 - Iliadis, Lazaros

A2 - Macintyre, John

A2 - Cortez, Paulo

PB - Springer

CY - Cham

ER -

Zai El Amri W, Tautz O, Ritter H, Melnik A. Transfer Learning with Jukebox for Music Source Separation. In Maglogiannis I, Iliadis L, Macintyre J, Cortez P, editors, Artificial Intelligence Applications and Innovations : 18th IFIP WG 12.5 International Conference, AIAI 2022, Hersonissos, Crete, Greece, June 17–20, 2022, Proceedings, Part II. 1. ed. Cham: Springer. 2022. p. 426-433. (IFIP Advances in Information and Communication Technology). doi: 10.1007/978-3-031-08337-2_35

Transfer Learning with Jukebox for Music Source Separation

Abstract

Publication series

Access to Document

Other files and links

Cite this