Can LLMs "Reason" in Music? An Evaluation of LLMs' Capability of Music Understanding and Generation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zhou, Ziya, Wu, Yuhang, Wu, Zhiyue, Zhang, Xinyue, Yuan, Ruibin, Ma, Yinghao, Wang, Lu, Benetos, Emmanouil, Xue, Wei, Guo, Yike
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910548601339904
author Zhou, Ziya
Wu, Yuhang
Wu, Zhiyue
Zhang, Xinyue
Yuan, Ruibin
Ma, Yinghao
Wang, Lu
Benetos, Emmanouil
Xue, Wei
Guo, Yike
author_facet Zhou, Ziya
Wu, Yuhang
Wu, Zhiyue
Zhang, Xinyue
Yuan, Ruibin
Ma, Yinghao
Wang, Lu
Benetos, Emmanouil
Xue, Wei
Guo, Yike
contents Symbolic Music, akin to language, can be encoded in discrete symbols. Recent research has extended the application of large language models (LLMs) such as GPT-4 and Llama2 to the symbolic music domain including understanding and generation. Yet scant research explores the details of how these LLMs perform on advanced music understanding and conditioned generation, especially from the multi-step reasoning perspective, which is a critical aspect in the conditioned, editable, and interactive human-computer co-creation process. This study conducts a thorough investigation of LLMs' capability and limitations in symbolic music processing. We identify that current LLMs exhibit poor performance in song-level multi-step music reasoning, and typically fail to leverage learned music knowledge when addressing complex musical tasks. An analysis of LLMs' responses highlights distinctly their pros and cons. Our findings suggest achieving advanced musical capability is not intrinsically obtained by LLMs, and future research should focus more on bridging the gap between music knowledge and reasoning, to improve the co-creation experience for musicians.
format Preprint
id arxiv_https___arxiv_org_abs_2407_21531
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Can LLMs "Reason" in Music? An Evaluation of LLMs' Capability of Music Understanding and Generation
Zhou, Ziya
Wu, Yuhang
Wu, Zhiyue
Zhang, Xinyue
Yuan, Ruibin
Ma, Yinghao
Wang, Lu
Benetos, Emmanouil
Xue, Wei
Guo, Yike
Sound
Computation and Language
Multimedia
Audio and Speech Processing
Symbolic Music, akin to language, can be encoded in discrete symbols. Recent research has extended the application of large language models (LLMs) such as GPT-4 and Llama2 to the symbolic music domain including understanding and generation. Yet scant research explores the details of how these LLMs perform on advanced music understanding and conditioned generation, especially from the multi-step reasoning perspective, which is a critical aspect in the conditioned, editable, and interactive human-computer co-creation process. This study conducts a thorough investigation of LLMs' capability and limitations in symbolic music processing. We identify that current LLMs exhibit poor performance in song-level multi-step music reasoning, and typically fail to leverage learned music knowledge when addressing complex musical tasks. An analysis of LLMs' responses highlights distinctly their pros and cons. Our findings suggest achieving advanced musical capability is not intrinsically obtained by LLMs, and future research should focus more on bridging the gap between music knowledge and reasoning, to improve the co-creation experience for musicians.
title Can LLMs "Reason" in Music? An Evaluation of LLMs' Capability of Music Understanding and Generation
topic Sound
Computation and Language
Multimedia
Audio and Speech Processing
url https://arxiv.org/abs/2407.21531