arXiv 2506.080652025-06astro-ph.IMcs.LG

天体物理观测反演中的动态扩散薛定谔桥

Dynamic Diffusion Schrödinger Bridge in Astrophysical Observational Inversions

LIX, École Polytechnique, IP Paris（巴黎高等理工学院LIX研究所）；Canadian Institute for Theoretical Astrophysics (CITA), University of Toronto（多伦多大学加拿大理论天体物理研究所）；Google DeepMind（谷歌DeepMind）；Department of Astronomy, University of Virginia（弗吉尼亚大学天文学系）；Department of Space, Earth & Environment, Chalmers University of Technology（楚科夫斯基技术大学空间、地球与环境系）

Ye Zhu, Duo Xu, Zhiwei Deng, Jonathan C. Tan, Olga Russakovsky

备注：: Accepted to NeurIPS 2025. Code available at https://github.com/L-YeZhu/AstroDSB. Updated funding information from last version

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

针对巨分子云恒星形成的观测逆预测任务，提出Astro-DSB模型（采用成对域假设的DSB变体），在模拟与真实数据中验证其优于传统方法，扩展扩散模型应用并为物理感知生成模型提供方向。

摘要：

我们在动态天体物理系统背景下研究扩散薛定谔桥（DSB）模型，具体解决巨分子云（GMCs）内恒星形成的观测逆预测任务。我们引入Astro-DSB模型，这是一种针对天体物理动力学、采用成对域假设的DSB变体。通过在物理模拟数据和真实观测数据（金牛座B213数据）中研究其学习过程和预测性能，我们得出两个主要结论。首先，从天体物理角度，我们提出的成对DSB方法在可解释性、学习效率和预测性能上优于传统天体统计方法和其他机器学习方法。其次，从生成建模角度，在具有未知初始条件和不同主导物理过程的物理模拟的分布外（OOD）测试案例中，概率生成建模相比判别式像素到像素建模表现出改进。我们的研究将扩散模型的研究扩展到传统视觉合成应用之外，并提供了模型学习能力超越纯数据统计的证据，为未来能使机器学习与真实（天体）物理系统动力学对齐的物理感知生成模型铺平了道路。

英文摘要：

We study Diffusion Schrödinger Bridge (DSB) models in the context of dynamical astrophysical systems, specifically tackling observational inverse prediction tasks within Giant Molecular Clouds (GMCs) for star formation. We introduce the Astro-DSB model, a variant of DSB with the pairwise domain assumption tailored for astrophysical dynamics. By investigating its learning process and prediction performance in both physically simulated data and in real observations (the Taurus B213 data), we present two main takeaways. First, from the astrophysical perspective, our proposed paired DSB method improves interpretability, learning efficiency, and prediction performance over conventional astrostatistical and other machine learning methods. Second, from the generative modeling perspective, probabilistic generative modeling reveals improvements over discriminative pixel-to-pixel modeling in Out-Of-Distribution (OOD) testing cases of physical simulations with unseen initial conditions and different dominant physical processes. Our study expands research into diffusion models beyond the traditional visual synthesis application and provides evidence of the models' learning abilities beyond pure data statistics, paving a path for future physics-aware generative models which can align dynamics between machine learning and real (astro)physical systems.

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。