Cover Song Identification in Practice with Multimodal Co-Training

Publikation: Beitrag in Buch/Bericht/Sammelwerk/KonferenzbandAufsatz in KonferenzbandForschungPeer-Review

Autoren

  • Simon Hachmeier
  • Robert Jäschke

Organisationseinheiten

Externe Organisationen

  • Humboldt-Universität zu Berlin (HU Berlin)
Forschungs-netzwerk anzeigen

Details

OriginalspracheEnglisch
Titel des SammelwerksLWDA 2023
UntertitelLernen, Wissen, Daten, Analysen 2023
Seiten359-371
Seitenumfang13
Band3630
PublikationsstatusVeröffentlicht - 2023
Veranstaltung2023 Learning. Knowledge. Data. Analytics, LWDA 2023 - Marburg, Deutschland
Dauer: 9 Okt. 202311 Okt. 2023

Publikationsreihe

NameCEUR Workshop Proceedings
Herausgeber (Verlag)CEUR Workshop Proceedings
Band3630
ISSN (Print)1613-0073

Abstract

The task of cover song identification (CSI) deals with the automatic matching of audio recordings by modeling musical similarity. CSI is of high relevance in the context of applications such as copyright infringement detection on online video platforms. Since online videos include metadata (eg. video titles, descriptions), one could leverage it for more effective CSI in practice. In this work, we experiment with state-of-the-art models of CSI and entity matching in a Co-Training ensemble. Our results outline slight improvements of the entity matching model. We further outline some suggestions for improvements of our approach to overcome the issue of overfitting CSI models which we observed.

ASJC Scopus Sachgebiete

Zitieren

Cover Song Identification in Practice with Multimodal Co-Training. / Hachmeier, Simon; Jäschke, Robert.
LWDA 2023: Lernen, Wissen, Daten, Analysen 2023. Band 3630 2023. S. 359-371 (CEUR Workshop Proceedings; Band 3630).

Publikation: Beitrag in Buch/Bericht/Sammelwerk/KonferenzbandAufsatz in KonferenzbandForschungPeer-Review

Hachmeier, S & Jäschke, R 2023, Cover Song Identification in Practice with Multimodal Co-Training. in LWDA 2023: Lernen, Wissen, Daten, Analysen 2023. Bd. 3630, CEUR Workshop Proceedings, Bd. 3630, S. 359-371, 2023 Learning. Knowledge. Data. Analytics, LWDA 2023, Marburg, Deutschland, 9 Okt. 2023. <https://ceur-ws.org/Vol-3630/LWDA2023-paper32.pdf>
Hachmeier, S., & Jäschke, R. (2023). Cover Song Identification in Practice with Multimodal Co-Training. In LWDA 2023: Lernen, Wissen, Daten, Analysen 2023 (Band 3630, S. 359-371). (CEUR Workshop Proceedings; Band 3630). https://ceur-ws.org/Vol-3630/LWDA2023-paper32.pdf
Hachmeier S, Jäschke R. Cover Song Identification in Practice with Multimodal Co-Training. in LWDA 2023: Lernen, Wissen, Daten, Analysen 2023. Band 3630. 2023. S. 359-371. (CEUR Workshop Proceedings).
Hachmeier, Simon ; Jäschke, Robert. / Cover Song Identification in Practice with Multimodal Co-Training. LWDA 2023: Lernen, Wissen, Daten, Analysen 2023. Band 3630 2023. S. 359-371 (CEUR Workshop Proceedings).
Download
@inproceedings{3bacc7d5513c4fce8ff0d24a22d740e0,
title = "Cover Song Identification in Practice with Multimodal Co-Training",
abstract = "The task of cover song identification (CSI) deals with the automatic matching of audio recordings by modeling musical similarity. CSI is of high relevance in the context of applications such as copyright infringement detection on online video platforms. Since online videos include metadata (eg. video titles, descriptions), one could leverage it for more effective CSI in practice. In this work, we experiment with state-of-the-art models of CSI and entity matching in a Co-Training ensemble. Our results outline slight improvements of the entity matching model. We further outline some suggestions for improvements of our approach to overcome the issue of overfitting CSI models which we observed.",
keywords = "co-training, cover song identification, entity matching",
author = "Simon Hachmeier and Robert J{\"a}schke",
year = "2023",
language = "English",
volume = "3630",
series = "CEUR Workshop Proceedings",
publisher = "CEUR Workshop Proceedings",
pages = "359--371",
booktitle = "LWDA 2023",
note = "2023 Learning. Knowledge. Data. Analytics, LWDA 2023 ; Conference date: 09-10-2023 Through 11-10-2023",

}

Download

TY - GEN

T1 - Cover Song Identification in Practice with Multimodal Co-Training

AU - Hachmeier, Simon

AU - Jäschke, Robert

PY - 2023

Y1 - 2023

N2 - The task of cover song identification (CSI) deals with the automatic matching of audio recordings by modeling musical similarity. CSI is of high relevance in the context of applications such as copyright infringement detection on online video platforms. Since online videos include metadata (eg. video titles, descriptions), one could leverage it for more effective CSI in practice. In this work, we experiment with state-of-the-art models of CSI and entity matching in a Co-Training ensemble. Our results outline slight improvements of the entity matching model. We further outline some suggestions for improvements of our approach to overcome the issue of overfitting CSI models which we observed.

AB - The task of cover song identification (CSI) deals with the automatic matching of audio recordings by modeling musical similarity. CSI is of high relevance in the context of applications such as copyright infringement detection on online video platforms. Since online videos include metadata (eg. video titles, descriptions), one could leverage it for more effective CSI in practice. In this work, we experiment with state-of-the-art models of CSI and entity matching in a Co-Training ensemble. Our results outline slight improvements of the entity matching model. We further outline some suggestions for improvements of our approach to overcome the issue of overfitting CSI models which we observed.

KW - co-training

KW - cover song identification

KW - entity matching

UR - http://www.scopus.com/inward/record.url?scp=85184666541&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:85184666541

VL - 3630

T3 - CEUR Workshop Proceedings

SP - 359

EP - 371

BT - LWDA 2023

T2 - 2023 Learning. Knowledge. Data. Analytics, LWDA 2023

Y2 - 9 October 2023 through 11 October 2023

ER -