Structure-informed Positional Encoding for Music Generation

Manvi Agarwal (1, 2) , Changhong Wang (1, 2) , Gaël Richard (1, 2)

1 S2A - Signal, Statistique et Apprentissage
2 IDS - Département Images, Données, Signal

Music generated by deep learning methods often suffers from a lack of coherence and long-term organization. Yet, multi-scale hierarchical structure is a distinctive feature of music signals. To leverage this information, we propose a structure-informed positional encoding framework for music generation with Transformers. We design three variants in terms of absolute, relative and non-stationary positional information. We comprehensively test them on two symbolic music generation tasks: next-timestep prediction and accompaniment generation. As a comparison, we choose multiple baselines from the literature and demonstrate the merits of our methods using several musically-motivated evaluation metrics. In particular, our methods improve the melodic and structural consistency of the generated pieces.

Mots clés: symbolic music generation; Transformers; music structure; positional encoding; symbolic music generation Transformers music structure positional encoding

Domaines: Son [cs.SD]; Intelligence artificielle [cs.AI]; Apprentissage [cs.LG]

Manvi Agarwal, Changhong Wang, Gaël Richard. Structure-informed Positional Encoding for Music Generation. IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Apr 2024, Seoul, South Korea. ⟨hal-04432659v3⟩