Imperfect score-matching leads to a shift between the training and the sampling distribution
of diffusion models. Due to the recursive nature of the generation process, errors in previous …

Diffusion models have demonstrated exceptional efficacy in various generative applications.
While existing models focus on minimizing a weighted sum of denoising score matching …

Removing reverb from reverberant music is a necessary technique to clean up audio for
downstream music manipulations. Reverberation of music contains two categories, natural …

Recent progress in deep generative models has improved the quality of neural vocoders in
speech domain. However, generating a high-quality singing voice remains challenging due …

In this paper we propose a novel generative approach, DiffRoll, to tackle automatic music
transcription (AMT). Instead of treating AMT as a discriminative task in which the model is …

The Fokker-Planck (FP) equation is a foundational PDE in stochastic processes. However,
curse of dimensionality (CoD) poses challenge when dealing with high-dimensional FP …

Although diffusion models (DMs) have shown promising performances in a number of tasks
(eg, speech synthesis and image generation), they might suffer from error propagation …

Denoising diffusion models have become ubiquitous for generative modeling. The core idea
is to transport the data distribution to a Gaussian by using a diffusion. Approximate samples …

We introduce an innovative approach for solving high-dimensional Fokker-Planck-L\'evy
(FPL) equations in modeling non-Brownian processes across disciplines such as physics …

While diffusion models have achieved promising performances in data synthesis, they might
suffer error propagation because of their cascade structure, where the distributional …