Improving Diffusion Models's Data-Corruption Resistance using Scheduled Pseudo-Huber Loss
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866916175452045312 |
|---|---|
| author | Khrapov, Artem Popov, Vadim Sadekova, Tasnima Yermekova, Assel Kudinov, Mikhail |
| author_facet | Khrapov, Artem Popov, Vadim Sadekova, Tasnima Yermekova, Assel Kudinov, Mikhail |
| contents | Diffusion models are known to be vulnerable to outliers in training data. In this paper we study an alternative diffusion loss function, which can preserve the high quality of generated data like the original squared $L_{2}$ loss while at the same time being robust to outliers. We propose to use pseudo-Huber loss function with a time-dependent parameter to allow for the trade-off between robustness on the most vulnerable early reverse-diffusion steps and fine details restoration on the final steps. We show that pseudo-Huber loss with the time-dependent parameter exhibits better performance on corrupted datasets in both image and audio domains. In addition, the loss function we propose can potentially help diffusion models to resist dataset corruption while not requiring data filtering or purification compared to conventional training algorithms. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2403_16728 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Improving Diffusion Models's Data-Corruption Resistance using Scheduled Pseudo-Huber Loss Khrapov, Artem Popov, Vadim Sadekova, Tasnima Yermekova, Assel Kudinov, Mikhail Artificial Intelligence Diffusion models are known to be vulnerable to outliers in training data. In this paper we study an alternative diffusion loss function, which can preserve the high quality of generated data like the original squared $L_{2}$ loss while at the same time being robust to outliers. We propose to use pseudo-Huber loss function with a time-dependent parameter to allow for the trade-off between robustness on the most vulnerable early reverse-diffusion steps and fine details restoration on the final steps. We show that pseudo-Huber loss with the time-dependent parameter exhibits better performance on corrupted datasets in both image and audio domains. In addition, the loss function we propose can potentially help diffusion models to resist dataset corruption while not requiring data filtering or purification compared to conventional training algorithms. |
| title | Improving Diffusion Models's Data-Corruption Resistance using Scheduled Pseudo-Huber Loss |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2403.16728 |