Pattern-Based Music Generation with Wasserstein Autoencoders and PRC Descriptions
Abstract
We demonstrate a pattern-based MIDI music generation system with a generation strategy based on Wasserstein autoencoders and a novel variant of pianoroll descriptions of patterns which employs separate channels for note velocities and note durations and can be fed into classic DCGAN-style convolutional architectures. We trained the system on two new datasets (in the acid-jazz and high-pop genres) composed by musicians in our team with music generation in mind. Our demonstration shows that moving smoothly in the latent space allows us to generate meaningful sequences of four-bars patterns.
Cite
Text
Borghuis et al. "Pattern-Based Music Generation with Wasserstein Autoencoders and PRC Descriptions." International Joint Conference on Artificial Intelligence, 2020. doi:10.24963/IJCAI.2020/751Markdown
[Borghuis et al. "Pattern-Based Music Generation with Wasserstein Autoencoders and PRC Descriptions." International Joint Conference on Artificial Intelligence, 2020.](https://mlanthology.org/ijcai/2020/borghuis2020ijcai-pattern/) doi:10.24963/IJCAI.2020/751BibTeX
@inproceedings{borghuis2020ijcai-pattern,
title = {{Pattern-Based Music Generation with Wasserstein Autoencoders and PRC Descriptions}},
author = {Borghuis, Valentijn and Angioloni, Luca and Brusci, Lorenzo and Frasconi, Paolo},
booktitle = {International Joint Conference on Artificial Intelligence},
year = {2020},
pages = {5225-5227},
doi = {10.24963/IJCAI.2020/751},
url = {https://mlanthology.org/ijcai/2020/borghuis2020ijcai-pattern/}
}