Unifying Bayesian Flow Networks and Diffusion Models through Stochastic Differential Equations
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2024
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866917681616125952 |
|---|---|
| author | Xue, Kaiwen Zhou, Yuhao Nie, Shen Min, Xu Zhang, Xiaolu Zhou, Jun Li, Chongxuan |
| author_facet | Xue, Kaiwen Zhou, Yuhao Nie, Shen Min, Xu Zhang, Xiaolu Zhou, Jun Li, Chongxuan |
| contents | Bayesian flow networks (BFNs) iteratively refine the parameters, instead of the samples in diffusion models (DMs), of distributions at various noise levels through Bayesian inference. Owing to its differentiable nature, BFNs are promising in modeling both continuous and discrete data, while simultaneously maintaining fast sampling capabilities. This paper aims to understand and enhance BFNs by connecting them with DMs through stochastic differential equations (SDEs). We identify the linear SDEs corresponding to the noise-addition processes in BFNs, demonstrate that BFN's regression losses are aligned with denoise score matching, and validate the sampler in BFN as a first-order solver for the respective reverse-time SDE. Based on these findings and existing recipes of fast sampling in DMs, we propose specialized solvers for BFNs that markedly surpass the original BFN sampler in terms of sample quality with a limited number of function evaluations (e.g., 10) on both image and text datasets. Notably, our best sampler achieves an increase in speed of 5~20 times for free. Our code is available at https://github.com/ML-GSAI/BFN-Solver. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2404_15766 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Unifying Bayesian Flow Networks and Diffusion Models through Stochastic Differential Equations Xue, Kaiwen Zhou, Yuhao Nie, Shen Min, Xu Zhang, Xiaolu Zhou, Jun Li, Chongxuan Machine Learning Artificial Intelligence Bayesian flow networks (BFNs) iteratively refine the parameters, instead of the samples in diffusion models (DMs), of distributions at various noise levels through Bayesian inference. Owing to its differentiable nature, BFNs are promising in modeling both continuous and discrete data, while simultaneously maintaining fast sampling capabilities. This paper aims to understand and enhance BFNs by connecting them with DMs through stochastic differential equations (SDEs). We identify the linear SDEs corresponding to the noise-addition processes in BFNs, demonstrate that BFN's regression losses are aligned with denoise score matching, and validate the sampler in BFN as a first-order solver for the respective reverse-time SDE. Based on these findings and existing recipes of fast sampling in DMs, we propose specialized solvers for BFNs that markedly surpass the original BFN sampler in terms of sample quality with a limited number of function evaluations (e.g., 10) on both image and text datasets. Notably, our best sampler achieves an increase in speed of 5~20 times for free. Our code is available at https://github.com/ML-GSAI/BFN-Solver. |
| title | Unifying Bayesian Flow Networks and Diffusion Models through Stochastic Differential Equations |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2404.15766 |