Nuclear physicists seek an accurate description of the properties of atomic nuclei, collisions between nuclei, and extreme environments such as the first few seconds of our universe or the interior of a neutron star. These situations involve many particles interacting through complex forces. They’re each described by a number of different models that typically explain accurately results of existing experiments. The models don’t do as well predicting what will happen in future experiments or in environments that are inaccessible here on Earth. The Bayesian Analysis of Nuclear Dynamics (BAND) Framework will use advanced statistical methods to produce forecasts for as-yet-unexplored situations that combine nuclear-physics models in an optimal way. These will be more reliable than the predictions of any individual model. BAND’s forefront computer codes will be widely available and will facilitate the design of nuclear-physics experiments that can deliver the largest gain in understanding. The adoption of BAND’s tools in other sciences dealing with “model uncertainty†could spur broad scientific innovation. Undergraduate and graduate students working on BAND will gain a broad range of technical skills in data science, machine learning, nuclear physics, and high-performance computing.
Nuclear physicists seek a quantitative description of strongly-interacting matter. Sophisticated models of how neutrons and protons interact in the nucleus, extreme environments, and collisions between nuclei have been key to the great progress made towards this goal. These models typically describe extant data well, but often yield divergent predictions for future experiments. The Bayesian Analysis of Nuclear Dynamics (BAND) framework will be a broadly available set of computational tools built through intensive collaboration between statisticians, computer scientists and nuclear physicists. It will combine the results of several models, incorporating prior knowledge and experimental data for each, to produce a full assessment of the uncertainty in nuclear-physics predictions. This will enable quantitative evaluation of the impact of future experiments, accelerating the theory-experiment feedback loop and spurring innovation. It will also help quantify uncertainties for terrestrially inaccessible environments, such as the core of neutron stars or the first microsecond after the Big Bang. Similar challenges are faced by researchers modeling complex dynamics in other sciences, so BAND’s tools will have broad appeal. Undergraduate and graduate students working on BAND will gain expertise in statistical methods and nuclear physics, as well as experience with large-scale computing and machine learning.
This award by the Office of Advanced Cyberinfrastructure is jointly supported by the Physics at the Information Frontier in the Division of Physics and the CDS&E program in the Division of Mathematical Sciences within the Directorate for Mathematical and Physical Sciences.
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.
