Score-based diffusion models have demonstrated remarkable empirical success in learning high-dimensional distributions, particularly those exhibiting low-dimensional and multi-modal structures. However, theoretical understanding of their statistical efficiency remains limited. Existing theories typically rely on strong regularity assumptions, such as uniformly bounded densities or globally smooth score functions, which fail to capture such intrinsic structures. In this work, we study the sample complexity of diffusion models for learning distributions supported on a union of low-dimensional subspaces. Assuming that the data distribution within each subspace is subgaussian, we show that diffusion models require at most $\widetilde{O}(\varepsilon^{-k \vee 2})$ samples to achieve $\varepsilon$ error in 1-Wasserstein distance, where $k$ is the intrinsic dimension. This near-optimal convergence rate depends only on the intrinsic dimension and significantly improves upon prior theoretical guarantees that suffer from the curse of dimensionality. Notably, our analysis applies to a broad collection of distributions without imposing smoothness, bounded-density, or log-concavity assumptions. Overall, our results show that diffusion models can statistically adapt to intrinsic low-dimensional structure while naturally accommodating multi-modal data, offering a rigorous theoretical justification for their success in complex high-dimensional learning tasks.
3M-Diffusion: Latent Multi-Modal Diffusion for Language-Guided Molecular Structure Generation
Artificial Intelligence Research Laboratory(人工智能研究实验室);College of Information Sciences and Technology(信息科学与技术学院);Center for Artificial Intelligence Foundations and Scientific Applications(人工智能基础与科学应用中心);Institute for Computational and Data Sciences(计算与数据科学研究所);The Pennsylvania State University(宾夕法尼亚州立大学)
Cross-modal Diffusion Modelling for Super-resolved Spatial Transcriptomics
Department of Clinical Neurosciences, University of Cambridge, UK(剑桥大学临床神经科学系);Department of Applied Mathematics and Theoretical Physics, University of Cambridge, UK(剑桥大学应用数学与理论物理系);School of Science and Engineering, University of Dundee, UK(邓迪大学科学与工程学院);School of Medicine, University of Dundee, UK(邓迪大学医学院);Zhejiang Lab, China(浙江实验室)
Cocktail: Mixing Multi-Modality Controls for Text-Conditional Image Generation
Nanyang Technological University(南洋理工大学);South China University of Technology(华南理工大学);University of Oxford(牛津大学);The University of Sydney(悉尼大学);JD Explore Academy(京东探索研究院)