Timezone: »

 
Score-based Denoising Diffusion with Non-Isotropic Gaussian Noise Models
Vikram Voleti · Chris Pal · Adam Oberman

@
in NeurIPS 2022 Workshop on Score-Based Methods »
Event URL: https://openreview.net/forum?id=igC8cJKcb0Q »

Generative models based on denoising diffusion techniques have led to an unprecedented increase in the quality and diversity of imagery that is now possible to create with neural generative models. However, most contemporary state-of-the-art methods are derived from a standard isotropic Gaussian formulation. In this work we examine the situation where non-isotropic Gaussian distributions are used. We present the key mathematical derivations for creating denoising diffusion models using an underlying non-isotropic Gaussian noise model. We also provide initial experiments to help verify empirically that this more general modelling approach can also yield high-quality samples.

Author Information

Vikram Voleti (Meta AI, Mila, University of Montreal)

I am a PhD candidate at Mila, University of Montreal, and a Research Intern at Meta AI. I work on generative models of images, 3D and video. My recent work is on score-based denoising diffusion model for video prediction, generation and interpolation.

Chris Pal (Montreal Institute for Learning Algorithms, École Polytechnique, Université de Montréal)
Adam Oberman (McGill University)

More from the Same Authors