Reward-Guided Iterative Refinement in Diffusion Models at Test-Time with Applications to Protein and DNA Design
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_ | 1866917988344528896 |
|---|---|
| author | Uehara, Masatoshi Su, Xingyu Zhao, Yulai Li, Xiner Regev, Aviv Ji, Shuiwang Levine, Sergey Biancalani, Tommaso |
| author_facet | Uehara, Masatoshi Su, Xingyu Zhao, Yulai Li, Xiner Regev, Aviv Ji, Shuiwang Levine, Sergey Biancalani, Tommaso |
| contents | To fully leverage the capabilities of diffusion models, we are often interested in optimizing downstream reward functions during inference. While numerous algorithms for reward-guided generation have been recently proposed due to their significance, current approaches predominantly focus on single-shot generation, transitioning from fully noised to denoised states. We propose a novel framework for inference-time reward optimization with diffusion models inspired by evolutionary algorithms. Our approach employs an iterative refinement process consisting of two steps in each iteration: noising and reward-guided denoising. This sequential refinement allows for the gradual correction of errors introduced during reward optimization. Besides, we provide a theoretical guarantee for our framework. Finally, we demonstrate its superior empirical performance in protein and cell-type-specific regulatory DNA design. The code is available at \href{https://github.com/masa-ue/ProDifEvo-Refinement}{https://github.com/masa-ue/ProDifEvo-Refinement}. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2502_14944 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Reward-Guided Iterative Refinement in Diffusion Models at Test-Time with Applications to Protein and DNA Design Uehara, Masatoshi Su, Xingyu Zhao, Yulai Li, Xiner Regev, Aviv Ji, Shuiwang Levine, Sergey Biancalani, Tommaso Machine Learning Artificial Intelligence Neural and Evolutionary Computing Quantitative Methods To fully leverage the capabilities of diffusion models, we are often interested in optimizing downstream reward functions during inference. While numerous algorithms for reward-guided generation have been recently proposed due to their significance, current approaches predominantly focus on single-shot generation, transitioning from fully noised to denoised states. We propose a novel framework for inference-time reward optimization with diffusion models inspired by evolutionary algorithms. Our approach employs an iterative refinement process consisting of two steps in each iteration: noising and reward-guided denoising. This sequential refinement allows for the gradual correction of errors introduced during reward optimization. Besides, we provide a theoretical guarantee for our framework. Finally, we demonstrate its superior empirical performance in protein and cell-type-specific regulatory DNA design. The code is available at \href{https://github.com/masa-ue/ProDifEvo-Refinement}{https://github.com/masa-ue/ProDifEvo-Refinement}. |
| title | Reward-Guided Iterative Refinement in Diffusion Models at Test-Time with Applications to Protein and DNA Design |
| topic | Machine Learning Artificial Intelligence Neural and Evolutionary Computing Quantitative Methods |
| url | https://arxiv.org/abs/2502.14944 |