Jiaying Yu. "Acoustic feature decoupling and pre-trained language model integration for music-to-text cross-modal generation." Scientific Reports (2026). https://doi.org/10.1038/s41598-026-58108-7