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

Autor:	Sehwag, Vikash, Hazirbas, Caner, Gordo, Albert, Ozgenel, Firat, Ferrer, Cristian Canton
Rok vydání:	2022
Předmět:	Computer Science - Computer Vision and Pattern Recognition Computer Science - Machine Learning
Druh dokumentu:	Working Paper
Popis:	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. Comment: CVPR 2022 (fixed some discrepancies in notation - v2)
Databáze:	arXiv
Externí odkaz:	http://arxiv.org/abs/2203.17260 Zobrazit plný text záznamu View this record from Arxiv