In this paper we present a new approach for the generation of
multi-instrument symbolic music driven by musical emotion. The principal
novelty of our approach centres on conditioning a state-of-the-art transformer
based on continuous-valued valence and arousal labels. In addition, we provide
a new large-scale dataset of symbolic music paired with emotion labels in terms
of valence and arousal. We evaluate our approach in a quantitative manner in
two ways, first by measuring its note prediction accuracy, and second via a
regression task in the valence-arousal plane. Our results demonstrate that our
proposed approaches outperform conditioning using control tokens which is
representative of the current state of the art.
Description
[2203.16165] Symbolic music generation conditioned on continuous-valued emotions
%0 Generic
%1 sulun2022symbolic
%A Sulun, Serkan
%A Davies, Matthew E. P.
%A Viana, Paula
%D 2022
%K emotions generation
%R 10.1109/ACCESS.2022.3169744
%T Symbolic music generation conditioned on continuous-valued emotions
%U http://arxiv.org/abs/2203.16165
%X In this paper we present a new approach for the generation of
multi-instrument symbolic music driven by musical emotion. The principal
novelty of our approach centres on conditioning a state-of-the-art transformer
based on continuous-valued valence and arousal labels. In addition, we provide
a new large-scale dataset of symbolic music paired with emotion labels in terms
of valence and arousal. We evaluate our approach in a quantitative manner in
two ways, first by measuring its note prediction accuracy, and second via a
regression task in the valence-arousal plane. Our results demonstrate that our
proposed approaches outperform conditioning using control tokens which is
representative of the current state of the art.
@misc{sulun2022symbolic,
abstract = {In this paper we present a new approach for the generation of
multi-instrument symbolic music driven by musical emotion. The principal
novelty of our approach centres on conditioning a state-of-the-art transformer
based on continuous-valued valence and arousal labels. In addition, we provide
a new large-scale dataset of symbolic music paired with emotion labels in terms
of valence and arousal. We evaluate our approach in a quantitative manner in
two ways, first by measuring its note prediction accuracy, and second via a
regression task in the valence-arousal plane. Our results demonstrate that our
proposed approaches outperform conditioning using control tokens which is
representative of the current state of the art.},
added-at = {2023-04-21T11:20:22.000+0200},
author = {Sulun, Serkan and Davies, Matthew E. P. and Viana, Paula},
biburl = {https://www.bibsonomy.org/bibtex/24377853878b15b0f50688276b20a76e4/alex_h},
description = {[2203.16165] Symbolic music generation conditioned on continuous-valued emotions},
doi = {10.1109/ACCESS.2022.3169744},
interhash = {c0c0406eb1fd0614a346f42079495ae8},
intrahash = {4377853878b15b0f50688276b20a76e4},
keywords = {emotions generation},
note = {cite arxiv:2203.16165Comment: Published in IEEE Access},
timestamp = {2023-04-21T11:20:22.000+0200},
title = {Symbolic music generation conditioned on continuous-valued emotions},
url = {http://arxiv.org/abs/2203.16165},
year = 2022
}