Ambient Dataloops: Generative Models for Dataset Refinement
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866918299378384896 |
|---|---|
| author | Rodríguez-Muñoz, Adrián Daspit, William Klivans, Adam Torralba, Antonio Daskalakis, Constantinos Daras, Giannis |
| author_facet | Rodríguez-Muñoz, Adrián Daspit, William Klivans, Adam Torralba, Antonio Daskalakis, Constantinos Daras, Giannis |
| contents | We propose Ambient Dataloops, an iterative framework for refining datasets that makes it easier for diffusion models to learn the underlying data distribution. Modern datasets contain samples of highly varying quality, and training directly on such heterogeneous data often yields suboptimal models. We propose a dataset-model co-evolution process; at each iteration of our method, the dataset becomes progressively higher quality, and the model improves accordingly. To avoid destructive self-consuming loops, at each generation, we treat the synthetically improved samples as noisy, but at a slightly lower noisy level than the previous iteration, and we use Ambient Diffusion techniques for learning under corruption. Empirically, Ambient Dataloops achieve state-of-the-art performance in unconditional and text-conditional image generation and de novo protein design. We further provide a theoretical justification for the proposed framework that captures the benefits of the data looping procedure. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_15417 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Ambient Dataloops: Generative Models for Dataset Refinement Rodríguez-Muñoz, Adrián Daspit, William Klivans, Adam Torralba, Antonio Daskalakis, Constantinos Daras, Giannis Machine Learning Artificial Intelligence I.4.3 We propose Ambient Dataloops, an iterative framework for refining datasets that makes it easier for diffusion models to learn the underlying data distribution. Modern datasets contain samples of highly varying quality, and training directly on such heterogeneous data often yields suboptimal models. We propose a dataset-model co-evolution process; at each iteration of our method, the dataset becomes progressively higher quality, and the model improves accordingly. To avoid destructive self-consuming loops, at each generation, we treat the synthetically improved samples as noisy, but at a slightly lower noisy level than the previous iteration, and we use Ambient Diffusion techniques for learning under corruption. Empirically, Ambient Dataloops achieve state-of-the-art performance in unconditional and text-conditional image generation and de novo protein design. We further provide a theoretical justification for the proposed framework that captures the benefits of the data looping procedure. |
| title | Ambient Dataloops: Generative Models for Dataset Refinement |
| topic | Machine Learning Artificial Intelligence I.4.3 |
| url | https://arxiv.org/abs/2601.15417 |