DeepAI AI Chat
Log In Sign Up

Generating High Fidelity Data from Low-density Regions using Diffusion Models

by   Vikash Sehwag, et al.
Princeton University

Our work focuses on addressing sample deficiency from low-density regions of data manifold in common image datasets. We leverage diffusion process based generative models to synthesize novel images from low-density regions. We observe that uniform sampling from diffusion models predominantly samples from high-density regions of the data manifold. Therefore, we modify the sampling process to guide it towards low-density regions while simultaneously maintaining the fidelity of synthetic data. We rigorously demonstrate that our process successfully generates novel high fidelity samples from low-density regions. We further examine generated samples and show that the model does not memorize low-density data and indeed learns to generate novel samples from low-density regions.


page 1

page 6

page 12

page 14

page 15

page 16

page 17

page 18


Don't Play Favorites: Minority Guidance for Diffusion Models

We explore the problem of generating minority samples using diffusion mo...

Generating High Fidelity Synthetic Data via Coreset selection and Entropic Regularization

Generative models have the ability to synthesize data points drawn from ...

Deep Equilibrium Approaches to Diffusion Models

Diffusion-based generative models are extremely effective in generating ...

Geometry-Based Data Generation

Many generative models attempt to replicate the density of their input d...

Latent Dynamical Implicit Diffusion Processes

Latent dynamical models are commonly used to learn the distribution of a...

Instance Selection for GANs

Recent advances in Generative Adversarial Networks (GANs) have led to th...

Msanii: High Fidelity Music Synthesis on a Shoestring Budget

In this paper, we present Msanii, a novel diffusion-based model for synt...