Symbolic Music Generation with Non-Differentiable Rule Guided Diffusion
This addresses the problem of generating high-quality, controllable symbolic music for applications like composition and entertainment, representing a novel technical contribution rather than an incremental improvement.
The paper tackles symbolic music generation by introducing Stochastic Control Guidance (SCG), a training-free method for non-differentiable rule guidance, and a latent diffusion architecture, resulting in marked advancements in music quality and controllability that outperform state-of-the-art generators.
We study the problem of symbolic music generation (e.g., generating piano rolls), with a technical focus on non-differentiable rule guidance. Musical rules are often expressed in symbolic form on note characteristics, such as note density or chord progression, many of which are non-differentiable which pose a challenge when using them for guided diffusion. We propose Stochastic Control Guidance (SCG), a novel guidance method that only requires forward evaluation of rule functions that can work with pre-trained diffusion models in a plug-and-play way, thus achieving training-free guidance for non-differentiable rules for the first time. Additionally, we introduce a latent diffusion architecture for symbolic music generation with high time resolution, which can be composed with SCG in a plug-and-play fashion. Compared to standard strong baselines in symbolic music generation, this framework demonstrates marked advancements in music quality and rule-based controllability, outperforming current state-of-the-art generators in a variety of settings. For detailed demonstrations, code and model checkpoints, please visit our project website: https://scg-rule-guided-music.github.io/.