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