2026-04-17T06:06:59Zhttps://recercat.cat/oai/request
oai:recercat.cat:10230/705432025-05-29T01:41:52Zcom_2072_6col_2072_452952
RECERCAT author Ramoneda, Pedro author Rocamora, Martín author Akama, Taketo 2025-05-29T01:41:52Z 2025-05-29T01:41:52Z 2025-05-28T06:07:26Z2025-05-28T06:07:26Z2024 http://hdl.handle.net/10230/70543 Autoregressive generative transformers are key in music generation, producing coherent compositions but facing challenges in human-machine collaboration. We propose RefinPaint, an iterative technique that improves the sampling process. It does this by identifying the weaker music elements using a feedback model, which then informs the choices for resampling by an inpainting model. This dualfocus methodology not only facilitates the machine’s ability to improve its automatic inpainting generation through repeated cycles but also offers a valuable tool for humans seeking to refine their compositions with automatic proofreading. Experimental results suggest RefinPaint’s effectiveness in inpainting and proofreading tasks, demonstrating its value for refining music created by both machines and humans. This approach not only facilitates creativity but also aids amateur composers in improving their work. © P. Ramoneda, M. Rocamora & T. Akama. Licensed under a Creative Commons Attribution 4.0 International License (CC BY 4.0). Attribution: P. Ramoneda, M. Rocamora & T. Akama, “Music Proofreading with RefinPaint: Where and How to Modify Compositions given Context”, in Proc. of the 25th Int. Society for Music Information Retrieval Conf., San Francisco, USA, 2024. http://creativecommons.org/licenses/by/4.0 info:eu-repo/semantics/openAccess Music proofreading RefinPaint Composition modifications Music proofreading with RefinPaint: where and how to modify compositions given context info:eu-repo/semantics/conferenceObject info:eu-repo/semantics/publishedVersion