Say What? Collaborative Pop Lyric Generation Using Multitask Transfer Learning

Naveen Ram; Tanay Gummadi; Rahul Bhethanabotla; Richard J. Savery; Gil Weinberg

doi:10.1145/3472307.3484175

Say What? Collaborative Pop Lyric Generation Using Multitask Transfer Learning

Naveen Ram
, Tanay Gummadi
, Rahul Bhethanabotla
, Richard J. Savery
, Gil Weinberg

Research output: A Conference proceeding or a Chapter in Book › Conference contribution › peer-review

10 Citations (Scopus)

Abstract

Lyric generation is a popular sub-field of natural language generation that has seen growth in recent years. Pop lyrics are of unique interest due to the genre's unique style and content, in addition to the high level of collaboration that goes on behind the scenes in the professional pop songwriting process. In this paper, we present a collaborative line-level lyric generation system that utilizes transfer-learning via the T5 transformer model, which, till date, has not been used to generate pop lyrics. By working and communicating directly with professional songwriters, we develop a model that is able to learn lyrical and stylistic tasks like rhyming, matching line beat requirements, and ending lines with specific target words. Our approach compares favorably to existing methods for multiple datasets and yields positive results from our online studies and interviews with industry songwriters.

Original language	English
Title of host publication	HAI 2021 - Proceedings of the 9th International User Modeling, Adaptation and Personalization Human-Agent Interaction
Editors	Kohei Ogawa, Tomoko Yonezawa , Gale M. Lucas, Hirotaka Osawa, Wafa Johal , Masahiro Shiomi
Publisher	Association for Computing Machinery, Inc
Pages	165-173
Number of pages	9
ISBN (Electronic)	9781450386203
DOIs	https://doi.org/10.1145/3472307.3484175
Publication status	Published - 9 Nov 2021
Externally published	Yes
Event	9th International User Modeling, Adaptation and Personalization Human-Agent Interaction, HAI 2021 - Virtual, Online, Japan Duration: 9 Nov 2021 → 11 Nov 2021

Publication series

Name	HAI 2021 - Proceedings of the 9th International User Modeling, Adaptation and Personalization Human-Agent Interaction

Conference

Conference	9th International User Modeling, Adaptation and Personalization Human-Agent Interaction, HAI 2021
Country/Territory	Japan
City	Virtual, Online
Period	9/11/21 → 11/11/21

Access to Document

10.1145/3472307.3484175

Cite this

Ram, N., Gummadi, T., Bhethanabotla, R., Savery, R. J., & Weinberg, G. (2021). Say What? Collaborative Pop Lyric Generation Using Multitask Transfer Learning. In K. Ogawa, T. Yonezawa , G. M. Lucas, H. Osawa, W. Johal , & M. Shiomi (Eds.), HAI 2021 - Proceedings of the 9th International User Modeling, Adaptation and Personalization Human-Agent Interaction (pp. 165-173). (HAI 2021 - Proceedings of the 9th International User Modeling, Adaptation and Personalization Human-Agent Interaction). Association for Computing Machinery, Inc. https://doi.org/10.1145/3472307.3484175

Ram, Naveen ; Gummadi, Tanay ; Bhethanabotla, Rahul et al. / Say What? Collaborative Pop Lyric Generation Using Multitask Transfer Learning. HAI 2021 - Proceedings of the 9th International User Modeling, Adaptation and Personalization Human-Agent Interaction. editor / Kohei Ogawa ; Tomoko Yonezawa ; Gale M. Lucas ; Hirotaka Osawa ; Wafa Johal ; Masahiro Shiomi. Association for Computing Machinery, Inc, 2021. pp. 165-173 (HAI 2021 - Proceedings of the 9th International User Modeling, Adaptation and Personalization Human-Agent Interaction).

@inproceedings{c1674e86fe2d4f96921a30a2161f6c04,

title = "Say What? Collaborative Pop Lyric Generation Using Multitask Transfer Learning",

abstract = "Lyric generation is a popular sub-field of natural language generation that has seen growth in recent years. Pop lyrics are of unique interest due to the genre's unique style and content, in addition to the high level of collaboration that goes on behind the scenes in the professional pop songwriting process. In this paper, we present a collaborative line-level lyric generation system that utilizes transfer-learning via the T5 transformer model, which, till date, has not been used to generate pop lyrics. By working and communicating directly with professional songwriters, we develop a model that is able to learn lyrical and stylistic tasks like rhyming, matching line beat requirements, and ending lines with specific target words. Our approach compares favorably to existing methods for multiple datasets and yields positive results from our online studies and interviews with industry songwriters.",

keywords = "collaborative AI, lyric generation, natural language generation, pop music, transfer learning, transformers",

author = "Naveen Ram and Tanay Gummadi and Rahul Bhethanabotla and Savery, \{Richard J.\} and Gil Weinberg",

note = "Publisher Copyright: {\textcopyright} 2021 ACM.; 9th International User Modeling, Adaptation and Personalization Human-Agent Interaction, HAI 2021 ; Conference date: 09-11-2021 Through 11-11-2021",

year = "2021",

month = nov,

day = "9",

doi = "10.1145/3472307.3484175",

language = "English",

series = "HAI 2021 - Proceedings of the 9th International User Modeling, Adaptation and Personalization Human-Agent Interaction",

publisher = "Association for Computing Machinery, Inc",

pages = "165--173",

editor = "Ogawa, \{Kohei \} and \{Yonezawa \}, \{Tomoko \} and \{M. Lucas\}, \{Gale \} and Osawa, \{Hirotaka \} and \{Johal \}, \{Wafa \} and Shiomi, \{Masahiro \}",

booktitle = "HAI 2021 - Proceedings of the 9th International User Modeling, Adaptation and Personalization Human-Agent Interaction",

}

Ram, N, Gummadi, T, Bhethanabotla, R, Savery, RJ & Weinberg, G 2021, Say What? Collaborative Pop Lyric Generation Using Multitask Transfer Learning. in K Ogawa, T Yonezawa , G M. Lucas, H Osawa, W Johal & M Shiomi (eds), HAI 2021 - Proceedings of the 9th International User Modeling, Adaptation and Personalization Human-Agent Interaction. HAI 2021 - Proceedings of the 9th International User Modeling, Adaptation and Personalization Human-Agent Interaction, Association for Computing Machinery, Inc, pp. 165-173, 9th International User Modeling, Adaptation and Personalization Human-Agent Interaction, HAI 2021, Virtual, Online, Japan, 9/11/21. https://doi.org/10.1145/3472307.3484175

Say What? Collaborative Pop Lyric Generation Using Multitask Transfer Learning. / Ram, Naveen; Gummadi, Tanay; Bhethanabotla, Rahul et al.
HAI 2021 - Proceedings of the 9th International User Modeling, Adaptation and Personalization Human-Agent Interaction. ed. / Kohei Ogawa; Tomoko Yonezawa ; Gale M. Lucas; Hirotaka Osawa; Wafa Johal ; Masahiro Shiomi. Association for Computing Machinery, Inc, 2021. p. 165-173 (HAI 2021 - Proceedings of the 9th International User Modeling, Adaptation and Personalization Human-Agent Interaction).

Research output: A Conference proceeding or a Chapter in Book › Conference contribution › peer-review

TY - GEN

T1 - Say What? Collaborative Pop Lyric Generation Using Multitask Transfer Learning

AU - Ram, Naveen

AU - Gummadi, Tanay

AU - Bhethanabotla, Rahul

AU - Savery, Richard J.

AU - Weinberg, Gil

PY - 2021/11/9

Y1 - 2021/11/9

N2 - Lyric generation is a popular sub-field of natural language generation that has seen growth in recent years. Pop lyrics are of unique interest due to the genre's unique style and content, in addition to the high level of collaboration that goes on behind the scenes in the professional pop songwriting process. In this paper, we present a collaborative line-level lyric generation system that utilizes transfer-learning via the T5 transformer model, which, till date, has not been used to generate pop lyrics. By working and communicating directly with professional songwriters, we develop a model that is able to learn lyrical and stylistic tasks like rhyming, matching line beat requirements, and ending lines with specific target words. Our approach compares favorably to existing methods for multiple datasets and yields positive results from our online studies and interviews with industry songwriters.

AB - Lyric generation is a popular sub-field of natural language generation that has seen growth in recent years. Pop lyrics are of unique interest due to the genre's unique style and content, in addition to the high level of collaboration that goes on behind the scenes in the professional pop songwriting process. In this paper, we present a collaborative line-level lyric generation system that utilizes transfer-learning via the T5 transformer model, which, till date, has not been used to generate pop lyrics. By working and communicating directly with professional songwriters, we develop a model that is able to learn lyrical and stylistic tasks like rhyming, matching line beat requirements, and ending lines with specific target words. Our approach compares favorably to existing methods for multiple datasets and yields positive results from our online studies and interviews with industry songwriters.

KW - collaborative AI

KW - lyric generation

KW - natural language generation

KW - pop music

KW - transfer learning

KW - transformers

UR - https://www.scopus.com/pages/publications/85119368640

U2 - 10.1145/3472307.3484175

DO - 10.1145/3472307.3484175

M3 - Conference contribution

AN - SCOPUS:85119368640

T3 - HAI 2021 - Proceedings of the 9th International User Modeling, Adaptation and Personalization Human-Agent Interaction

SP - 165

EP - 173

BT - HAI 2021 - Proceedings of the 9th International User Modeling, Adaptation and Personalization Human-Agent Interaction

A2 - Ogawa, Kohei

A2 - Yonezawa , Tomoko

A2 - M. Lucas, Gale

A2 - Osawa, Hirotaka

A2 - Johal , Wafa

A2 - Shiomi, Masahiro

PB - Association for Computing Machinery, Inc

T2 - 9th International User Modeling, Adaptation and Personalization Human-Agent Interaction, HAI 2021

Y2 - 9 November 2021 through 11 November 2021

ER -

Ram N, Gummadi T, Bhethanabotla R, Savery RJ, Weinberg G. Say What? Collaborative Pop Lyric Generation Using Multitask Transfer Learning. In Ogawa K, Yonezawa T, M. Lucas G, Osawa H, Johal W, Shiomi M, editors, HAI 2021 - Proceedings of the 9th International User Modeling, Adaptation and Personalization Human-Agent Interaction. Association for Computing Machinery, Inc. 2021. p. 165-173. (HAI 2021 - Proceedings of the 9th International User Modeling, Adaptation and Personalization Human-Agent Interaction). doi: 10.1145/3472307.3484175

Say What? Collaborative Pop Lyric Generation Using Multitask Transfer Learning

Abstract

Publication series

Conference

Access to Document

Other files and links

Fingerprint

Cite this