让音乐生成模型更懂人类审美偏好，提升创作真实感与体验感。
Aligning Generative Music AI with Human Preferences: Methods and Challenges
- 引入偏好对齐技术，改进音乐生成的优化目标
- 解决时序连贯性、和声一致性等核心难题
- 适合音乐科技、人机共创领域研究者参考
生成式音乐AI在音质保真度和风格多样性上取得显著进展，但因使用特定损失函数，常与人类复杂审美偏好脱节。本文倡导系统性应用偏好对齐技术于音乐生成，弥合计算优化与人类听觉审美的根本差距。基于MusicRL的大规模偏好学习、DiffRhythm+中的多偏好对齐框架及Text2midi-InferAlign的推理阶段优化技术，探讨其如何应对音乐生成中特有的挑战：时序连贯性、和声一致性与主观质量评估。识别出关键研究挑战，包括长篇作品的可扩展性与偏好建模的可靠性问题。展望未来，偏好对齐的音乐生成有望推动交互式作曲工具与个性化音乐服务的发展。本文呼吁结合机器学习与音乐理论的跨学科研究，构建真正满足人类创造与体验需求的音乐AI系统。
原文摘要 · Abstract (English)
Recent advances in generative AI for music have achieved remarkable fidelity and stylistic diversity, yet these systems often fail to align with nuanced human preferences due to the specific loss functions they use. This paper advocates for the systematic application of preference alignment techniques to music generation, addressing the fundamental gap between computational optimization and human musical appreciation. Drawing on recent breakthroughs including MusicRL's large-scale preference learning, multi-preference alignment frameworks like diffusion-based preference optimization in DiffRhythm+, and inference-time optimization techniques like Text2midi-InferAlign, we discuss how these techniques can address music's unique challenges: temporal coherence, harmonic consistency, and subjective quality assessment. We identify key research challenges including scalability to long-form compositions, reliability amongst others in preference modelling. Looking forward, we envision preference-aligned music generation enabling transformative applications in interactive composition tools and personalized music services. This work calls for sustained interdisciplinary research combining advances in machine learning, music-theory to create music AI systems that truly serve human creative and experiential needs.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。