Disentangling Total-Variance and Signal-to-Noise-Ratio Improves Diffusion Models
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866913869284245504 |
|---|---|
| author | Kahouli, Khaled Ripken, Winfried Gugler, Stefan Unke, Oliver T. Müller, Klaus-Robert Nakajima, Shinichi |
| author_facet | Kahouli, Khaled Ripken, Winfried Gugler, Stefan Unke, Oliver T. Müller, Klaus-Robert Nakajima, Shinichi |
| contents | The long sampling time of diffusion models remains a significant bottleneck, which can be mitigated by reducing the number of diffusion time steps. However, the quality of samples with fewer steps is highly dependent on the noise schedule, i.e., the specific manner in which noise is introduced and the signal is reduced at each step. Although prior work has improved upon the original variance-preserving and variance-exploding schedules, these approaches $\textit{passively}$ adjust the total variance, without direct control over it. In this work, we propose a novel total-variance/signal-to-noise-ratio disentangled (TV/SNR) framework, where TV and SNR can be controlled independently. Our approach reveals that schedules where the TV explodes exponentially can often be improved by adopting a constant TV schedule while preserving the same SNR schedule. Furthermore, generalizing the SNR schedule of the optimal transport flow matching significantly improves the generation performance. Our findings hold across various reverse diffusion solvers and diverse applications, including molecular structure and image generation. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2502_08598 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Disentangling Total-Variance and Signal-to-Noise-Ratio Improves Diffusion Models Kahouli, Khaled Ripken, Winfried Gugler, Stefan Unke, Oliver T. Müller, Klaus-Robert Nakajima, Shinichi Machine Learning The long sampling time of diffusion models remains a significant bottleneck, which can be mitigated by reducing the number of diffusion time steps. However, the quality of samples with fewer steps is highly dependent on the noise schedule, i.e., the specific manner in which noise is introduced and the signal is reduced at each step. Although prior work has improved upon the original variance-preserving and variance-exploding schedules, these approaches $\textit{passively}$ adjust the total variance, without direct control over it. In this work, we propose a novel total-variance/signal-to-noise-ratio disentangled (TV/SNR) framework, where TV and SNR can be controlled independently. Our approach reveals that schedules where the TV explodes exponentially can often be improved by adopting a constant TV schedule while preserving the same SNR schedule. Furthermore, generalizing the SNR schedule of the optimal transport flow matching significantly improves the generation performance. Our findings hold across various reverse diffusion solvers and diverse applications, including molecular structure and image generation. |
| title | Disentangling Total-Variance and Signal-to-Noise-Ratio Improves Diffusion Models |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2502.08598 |