arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.00960cs.LGcs.AIcs.CEstat.COstat.ML
收藏

多模态科学学习超越扩散与流

Multimodal Scientific Learning Beyond Diffusions and Flows

Graduate Group in Applied Mathematics and Computational Science(应用数学与计算科学联合研究生组);Department of Computer and Information Science(计算机与信息科学系);Department of Mechanical Engineering and Applied Mechanics(机械工程与应用力学系)

Leonardo Ferreira Guilhoto, Akshat Kaushal, Paris Perdikaris

AI总结:

本文提出混合密度网络作为多模态科学学习中更高效、更稳定的替代方法,实现对科学问题中多模态不确定性的有效建模与解决。

摘要:

科学机器学习(SciML)日益需要能够捕捉来自不明确反问题、多稳定性以及混沌动态中多模态条件不确定性的模型。尽管近期工作倾向于高度表达的隐式生成模型,如扩散和基于流的方法,但这些方法往往数据需求高、计算成本大,并且与科学问题中常见的结构化解空间不匹配。我们证明混合密度网络(MDNs)为多模态不确定性量化提供了系统而被忽视的替代方案。作为显式参数密度估计器,MDNs施加了针对低维多模态物理定制的归纳偏差,使概率质量能够直接在全球范围内分配到不同的解分支中。这种结构在科学数据稀缺的领域中实现了强大的数据效率,能够可靠地恢复分离的模式。我们通过统一的概率框架对比显式和隐式分布网络,证明MDNs在一系列反问题、多稳定性以及混沌科学回归任务中实现了更优的泛化能力、可解释性和样本效率。

英文摘要:

Scientific machine learning (SciML) increasingly requires models that capture multimodal conditional uncertainty arising from ill-posed inverse problems, multistability, and chaotic dynamics. While recent work has favored highly expressive implicit generative models such as diffusion and flow-based methods, these approaches are often data-hungry, computationally costly, and misaligned with the structured solution spaces frequently found in scientific problems. We demonstrate that Mixture Density Networks (MDNs) provide a principled yet largely overlooked alternative for multimodal uncertainty quantification in SciML. As explicit parametric density estimators, MDNs impose an inductive bias tailored to low-dimensional, multimodal physics, enabling direct global allocation of probability mass across distinct solution branches. This structure delivers strong data efficiency, allowing reliable recovery of separated modes in regimes where scientific data is scarce. We formalize these insights through a unified probabilistic framework contrasting explicit and implicit distribution networks, and demonstrate empirically that MDNs achieve superior generalization, interpretability, and sample efficiency across a range of inverse, multistable, and chaotic scientific regression tasks.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑