Abstract
Jamming requires coordination, anticipation, and collaborative creativity between musicians. Current generative models of music produce expressive output but are not able to generate in an \emph{online} manner, meaning simultaneously with other musicians (human or otherwise). We propose ReaLchords, an online generative model for improvising chord accompaniment to user melody. We start with an online model pretrained by maximum likelihood, and use reinforcement learning to finetune the model for online use. The finetuning objective leverages both a novel reward model that provides feedback on both harmonic and temporal coherency between melody and chord, and a divergence term that implements a novel type of distillation from a teacher model that can see the future melody. Through quantitative experiments and listening tests, we demonstrate that the resulting model adapts well to unfamiliar input and produce fitting accompaniment. ReaLchords opens the door to live jamming, as well as simultaneous co-creation in other modalities.
Abstract (translated)
即兴演奏要求音乐家之间进行协调、预判和协作性创作。目前的音乐生成模型虽然能够产出有表现力的作品，但它们无法在线（即实时地）与其他音乐家同步生成音乐。我们提出了ReaLchords这一在线生成模型，旨在为用户提供的旋律即兴伴奏和弦。 我们的方法首先使用最大似然预训练一个在线模型，并通过强化学习对这个模型进行微调以适应在线应用的需求。在微调过程中，我们引入了一个新颖的奖励模型来评估旋律与和弦之间的和谐性和时间一致性，同时还有一个散度项，它从一个能够预见未来旋律的教师模型中提取知识（实施了一种新的蒸馏方法）。通过定量实验和听觉测试，我们证明了该模型在处理不熟悉的输入时表现良好，并能生成合适的伴奏。 ReaLchords不仅为实时即兴演奏打开了大门，还使得其他模态下的同时创作成为可能。
URL
https://arxiv.org/abs/2506.14723
