Advances in machine learning methods provide tools that have broad applicability in scientific research. These techniques are being applied across the diversity of nuclear physics research topics, leading to advances that will facilitate scientific discoveries and societal applications. This Colloquium provides a snapshot of nuclear physics research, which has been transformed by machine learning techniques.
The absolute scale of the neutrino mass plays a critical role in physics at every scale, from the subatomic to the cosmological. Measurements of the tritium end-point spectrum have provided the most precise direct limit on the neutrino mass scale. In this Letter, we present advances by Project 8 to the cyclotron radiation emission spectroscopy (CRES) technique culminating in the first frequency-based neutrino mass limit. With only a cm^{3}-scale physical detection volume, a limit of m_{β}<155 eV/c^{2} (152 eV/c^{2}) is extracted from the background-free measurement of the continuous tritium beta spectrum in a Bayesian (frequentist) analysis. Using ^{83m}Kr calibration data, a resolution of 1.66±0.19 eV (FWHM) is measured, the detector response model is validated, and the efficiency is characterized over the multi-keV tritium analysis window. These measurements establish the potential of CRES for a high-sensitivity next-generation direct neutrino mass experiment featuring low background and high resolution.
18 October 2021 in Physical Review Accelerators and Beams
We describe a method for precisely regulating the gradient magnet power supply (GMPS) at the Fermilab Booster accelerator complex using a neural network trained via reinforcement learning. We demonstrate preliminary results by training a surrogate machine-learning model on real accelerator data to emulate the GMPS, and using this surrogate model in turn to train the neural network for its regulation task. We additionally show how the neural networks to be deployed for control purposes may be compiled to execute on field-programmable gate arrays (FPGAs), and show the first machine-learning based control algorithm implemented on an FPGA for controls at the Fermilab accelerator complex. As there are no surprise latencies on an FPGA, this capability is important for operational stability in complicated environments such as an accelerator facility.
15 December 2022 in Physical Review Accelerators and Beams
High-power particle accelerators are complex machines with thousands of pieces of equipment that are frequently running at the cutting edge of technology. In order to improve the day-to-day operations and maximize the delivery of the science, new analytical techniques are being explored for anomaly detection, classification, and prognostications. As such, we describe the application of an uncertainty aware Machine Learning method using the Siamese neural network model to predict upcoming errant beam pulses using the data from a single monitoring device. By predicting the upcoming failure, we can stop the accelerator before damage occurs. We describe the accelerator operation, related Machine Learning research, the prediction performance required to abort the beam while maintaining operations, the monitoring device and its data, and the uncertainty aware Siamese method and its results. These results show that the researched method can be applied to improve accelerator operations.
Project 8 has developed a novel technique, cyclotron radiation emission spectroscopy (CRES), for direct neutrino mass measurements. A CRES-based experiment on the beta spectrum of tritium has been carried out in a small-volume apparatus. We provide a detailed account of the experiment, focusing on systematic effects and analysis techniques. In a Bayesian (frequentist) analysis, we measure the tritium endpoint as () eV and set upper limits of 155 (152) eV (90% C.L.) on the neutrino mass. No background events are observed beyond the endpoint in 82 days of running. We also demonstrate an energy resolution of in a resolution-optimized magnetic trap configuration by measuring -keV internal-conversion electrons. These measurements establish CRES as a low-background, high-resolution technique with the potential to advance neutrino mass sensitivity.