Collaborative Research: Science-Aware Computational Methods for Accelerating Data-Intensive Discovery: Astroparticle Physics as a Test Case

过去二十年的快速技术进步已经为几个学科开创了一个数据丰富的科学时代。其中一门学科是天体粒子物理学，研究人员的目标是通过直接探测暗物质来发现我们的宇宙是由什么组成的。如果使用数据科学工具从数据中提取重要的领域特定信息，并可靠地大规模测试科学假设，则可以加速这一发现。这个为期两年的项目的总体目标是为在科学学科背景下将科学知识融入机器学习和数据科学方法奠定基础，在这些学科背景下，发现需要对多个不完美传感器收集的大量噪声数据进行有效、高效的分析。在这样做的过程中，它不仅推动了数据科学、机器学习和天体物理学的最新发展，而且还有可能加速数据具有类似特征的其他科学学科中数据驱动的发现。该项目将开发基于概率图模型和图正则化逆问题的创新领域增强数据科学方法。使用领先的天体粒子实验XENON作为测试平台，研究人员将探索和演示将领域知识纳入机器学习和数据科学方法的方法。在此过程中，研究人员将解决暗物质识别背景下的主要数据分析挑战。此外，研究人员将投入大量精力与其他数据密集型科学社区接触，如材料科学、海洋学和气象学，这些社区可以从新的方法和想法中受益。该项目是美国国家科学基金会“利用数据革命（HDR）大创意”活动的一部分。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

The rapid technological advances of the last two decades have ushered in an era of data-rich science for several disciplines. One such discipline is astroparticle physics, where researchers aim to discover what our Universe is made of by trying to directly detect Dark Matter. This discovery can be hastened if data science tools are used to extract significant domain-specific information from data, and to reliably test scientific hypotheses at scale. The overarching goal of this two-year project is to lay the groundwork for incorporating scientific knowledge into machine learning and data science methods in the context of scientific disciplines in which discovery requires effective, efficient analysis of lots of noisy data gathered by multiple imperfect sensors. In doing so, it not only advances the state-of-the-art in data science, machine learning, and astrophysics, but it also has the potential to accelerate data-driven discoveries in other scientific disciplines where data shares similar characteristics.This project will develop innovative domain-enhanced data science methods that will be based on probabilistic graphical models and graph-regularized inverse problems. Using the leading astroparticle experiment XENON as a test bed, the investigators will explore and demonstrate approaches for incorporating domain knowledge into machine learning and data science methods. In doing so, the investigators will address major data-analysis challenges in the context of dark matter identification. Additionally, the investigators will invest significant effort reaching out to other data-intensive science communities, such as materials science, oceanography, and meteorology, that can benefit from the new methods and ideas. This project is part of the National Science Foundation's Harnessing the Data Revolution (HDR) Big Idea activity.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.