Understanding reinforcement learning-based fine-tuning of diffusion models: A tutorial and review

M Uehara, Y Zhao, T Biancalani, S Levine - arXiv preprint arXiv …, 2024 - arxiv.org
This tutorial provides a comprehensive survey of methods for fine-tuning diffusion models to
optimize downstream reward functions. While diffusion models are widely known to provide …

Decoding biology with massively parallel reporter assays and machine learning

A La Fleur, Y Shi, G Seelig - Genes & Development, 2024 - genesdev.cshlp.org
Massively parallel reporter assays (MPRAs) are powerful tools for quantifying the impacts of
sequence variation on gene expression. Reading out molecular phenotypes with …

Derivative-free guidance in continuous and discrete diffusion models with soft value-based decoding

X Li, Y Zhao, C Wang, G Scalia, G Eraslan… - arXiv preprint arXiv …, 2024 - arxiv.org
Diffusion models excel at capturing the natural design spaces of images, molecules, DNA,
RNA, and protein sequences. However, rather than merely generating designs that are …

Fine-tuning discrete diffusion models via reward optimization with applications to dna and protein design

C Wang, M Uehara, Y He, A Wang, T Biancalani… - arXiv preprint arXiv …, 2024 - arxiv.org
Recent studies have demonstrated the strong empirical performance of diffusion models on
discrete sequences across domains from natural language to biological sequence …

Constrained Diffusion Models via Dual Training

S Khalafi, D Ding, A Ribeiro - arXiv preprint arXiv:2408.15094, 2024 - arxiv.org
Diffusion models have attained prominence for their ability to synthesize a probability
distribution for a given dataset via a diffusion process, enabling the generation of new data …

Stochastic Control for Fine-tuning Diffusion Models: Optimality, Regularity, and Convergence

Y Han, M Razaviyayn, R Xu - arXiv preprint arXiv:2412.18164, 2024 - arxiv.org
Diffusion models have emerged as powerful tools for generative modeling, demonstrating
exceptional capability in capturing target data distributions from large datasets. However …

Maximum Entropy Inverse Reinforcement Learning of Diffusion Models with Energy-Based Models

S Yoon, H Hwang, D Kwon, YK Noh… - arXiv preprint arXiv …, 2024 - arxiv.org
We present a maximum entropy inverse reinforcement learning (IRL) approach for improving
the sample quality of diffusion generative models, especially when the number of generation …

Generative bandit optimization via diffusion posterior sampling

R De Santi, N Liniger, A Krause - openreview.net
Many real-world discovery problems, including drug and material design, can be modeled
within the bandit optimization framework, where an agent selects a sequence of experiments …