首次在函数空间中构建概率流常微分方程,加速无限维扩散模型推理。
Probability-Flow ODE in Infinite-Dimensional Function Spaces
- 在无限维函数空间中推导出概率流常微分方程的新形式。
- 减少函数求值次数,保持生成样本质量,适用于偏微分方程等任务。
- 适合研究无限维生成模型、物理信息机器学习的学者参考。
近年来,无限维扩散模型在本质为无限维结构的函数生成任务中展现出高效性和可扩展性。为加速此类模型的推理过程,本文首次在无限维函数空间中推导出概率流常微分方程(PF-ODE)的对应形式。基于这一新构建的PF-ODE,我们显著减少了函数求值次数,同时在函数生成任务中保持了高质量样本输出,包括对偏微分方程(PDEs)的应用。该方法为高维函数建模提供了更高效的计算路径。
原文摘要 · Abstract (English)
Recent advances in infinite-dimensional diffusion models have demonstrated their effectiveness and scalability in function generation tasks where the underlying structure is inherently infinite-dimensional. To accelerate inference in such models, we derive, for the first time, an analog of the probability-flow ODE (PF-ODE) in infinite-dimensional function spaces. Leveraging this newly formulated PF-ODE, we reduce the number of function evaluations while maintaining sample quality in function generation tasks, including applications to PDEs.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品;中文卡片由大模型生成,请以原文为准。