Accelerating Exploration in Plasma and Radiation Physics using Bayesian Optimization

Modern scientific and physical experiments are frequently complex and multidimensional in nature, especially with the recent big data explosion. However, the exploration of such a large hyperspace to identify the optimal solution is still guided by methods based on experience and intuition similar to those applied for simpler one-dimensional experiments. Unfortunately, despite the recent advances in computing and machine learning, optimization tools to accelerate and guide experimental design are not widely understood and therefore underutilized. In this work, we develop and explore a novel Bayesian optimization methodology and apply it to navigate the large multidimensional space of modern plasma and radiation physics to adaptively design the next experiment that will bring us closer to the optimal solution. The foundations, construction, and methods of Bayesian Optimization are provided in this work. We further demonstrate the potential use of Bayesian optimization by (1) guiding parameter optimization for atmospheric pressure, pulsed plasmas for inactivating microorganisms and (2) accelerating the detection of a hidden radiation source to enhance emergency response and nuclear nonproliferation capabilities. Even with limited data, our approach appears to successfully identify the parameters needed to perform the next experiment that will move the solution closer to the optimal. We show that after only a few iterative steps, the algorithm can correctly reach the optimal solution and accelerate the experimental process, while decreasing cost and time.