{"type":"ScholarlyArticle","title":"QA-MDT: Quality-aware Masked Diffusion Transformer for Enhanced Music Generation","authors":[{"name":"Chang Li","slug":"chang-li","given":"Chang","family":"Li","affiliation":""},{"name":"Ruoyu Wang","slug":"ruoyu-wang","given":"Ruoyu","family":"Wang","affiliation":""},{"name":"Lijuan Liu","slug":"lijuan-liu","given":"Lijuan","family":"Liu","affiliation":""},{"name":"Jun Du","slug":"jun-du","given":"Jun","family":"Du","affiliation":""},{"name":"Yixuan Sun","slug":"yixuan-sun","given":"Yixuan","family":"Sun","affiliation":""},{"name":"Zilu Guo","slug":"zilu-guo","given":"Zilu","family":"Guo","affiliation":""},{"name":"Zhengrong Zhang","slug":"zhengrong-zhang","given":"Zhengrong","family":"Zhang","affiliation":""},{"name":"Yuan Jiang","slug":"yuan-jiang","given":"Yuan","family":"Jiang","affiliation":""},{"name":"Jianqing Gao","slug":"jianqing-gao","given":"Jianqing","family":"Gao","affiliation":""},{"name":"Feng Ma","slug":"feng-ma","given":"Feng","family":"Ma","affiliation":""}],"abstract":"Text-to-music (TTM) generation, which converts textual descriptions into audio, opens up innovative avenues for multimedia creation.\r\nAchieving high quality and diversity in this process demands extensive, high-quality data, which are often scarce in available datasets. Most open-source datasets frequently suffer from issues like low-quality waveforms and low text-audio consistency, hindering the advancement of music generation models.\r\nTo address these challenges, we propose a novel quality-aware training paradigm for generating high-quality, high-musicality music from large-scale, quality-imbalanced datasets. Additionally, by leveraging unique properties in the latent space of musical signals, we adapt and implement a masked diffusion transformer (MDT) model for the TTM task, showcasing its capacity for quality control and enhanced musicality. Furthermore, we introduce a three-stage caption refinement approach to address low-quality captions' issue. Experiments show state-of-the-art (SOTA) performance on benchmark datasets including MusicCaps and the Song-Describer Dataset with both objective and subjective metrics.\r\nDemo audio samples are available at https://qa-mdt.github.io/, code and pretrained checkpoints are open-sourced at https://github.com/ivcylc/OpenMusic.","doi":"10.24963/ijcai.2025/1126","url":"https://ijcai.org/proceedings/2025/1126","pdf":"https://ijcai.org/proceedings/2025/1126.pdf","markdown":"https://ijcai.org/proceedings/2025/1126.md","bibtex":"https://ijcai.org/proceedings/2025/bibtex/1126","volume":{"slug":"2025","title":"Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence","short_title":"IJCAI 2025","url":"https://ijcai.org/proceedings/2025/","year":2025,"location":"Montreal, Canada","dates":"16-22 August 2025","editors":"James Kwok","isbn":"978-1-956792-06-5","publisher":"International Joint Conferences on Artificial Intelligence Organization","doi_prefix":"10.24963/ijcai.2025"},"section":"AI, Arts & Creativity","track":"AI, Arts & Creativity","number":1126,"pages":{"first":10135,"last":10143},"keywords":["Application domains: General","Application domains: Music and sound","Theory and philosophy of arts and creativity in AI systems: Autonomous creative or artistic AI","Methods and resources: Machine learning, deep learning, neural models, reinforcement learning"],"videos":[],"references":[],"license_note":"Copyright International Joint Conferences on Artificial Intelligence."}