Generating High Fidelity Data from Low-density Regions using Diffusion Models
- Resource Type
- Conference
- Authors
- Sehwag, Vikash; Hazirbas, Caner; Gordo, Albert; Ozgenel, Firat; Ferrer, Cristian Canton
- Source
- 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) CVPR Computer Vision and Pattern Recognition (CVPR), 2022 IEEE/CVF Conference on. :11482-11491 Jun, 2022
- Subject
- Computing and Processing
Manifolds
Computer vision
Computational modeling
Diffusion processes
Data models
Pattern recognition
Image and video synthesis and generation; Machine learning; Representation learning
- Language
- ISSN
- 2575-7075
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 simulta-neously 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.