| Summary: | 2nd Place at Denman Undergraduate Research Forum In this project we have investigated how genetic programming, a form a machine learning, can improve the analysis of data from the Antarctic Impulsive Transient Antenna (ANITA), a balloon experiment searching for ultra-high-energy (UHE) neutrinos that interact in the ice in Antarctica. Discovering these UHE neutrinos will unlock new information about the universe¬¬ and will lead the way into a new era of neutrino astronomy. ANITA, like many astroparticle experiments, relies heavily on being able to differentiate between signal events and background noise. This project has taken advantage of genetic programming algorithms to effectively model the anthropogenic backgrounds. Genetic programming takes advantage of an evolutionary style of function generation, in which prospective functions meant to describe a dataset are populated and tested in sets called “generations”, with the best fitting functions populating the next generation of functions. One such program that implements a genetic programming algorithm is called Karoo GP, a program written by Kai Staats, a scientist currently working with the Laser Interferometer Gravitational-Wave Observatory (LIGO). Karoo GP was designed for the Square Kilometer Array (SKA) radio experiment and has been used by the LIGO Collaboration. Karoo GP outputs the fitting algorithms as analytical functions, which would easily allow us to include it in the analysis. In this research project, Karoo GP was used to model the data from the ANITA experiment. Events in the data were characterized by two variables, the signal-to-noise ratio (SNR) and the correlation between the voltage peaks from different antennas. At the moment, the best model from Karoo GP does not adequately describe the data, and will not be used in further analysis. However, we suggest ways in which the algorithm can be improved, and it is possible that Karoo GP will allow for better optimization of the neutrino flux limit in the future. NSF CAREER Award 125557 ANITA Grant NNX15AC20G College of Engineering No embargo Academic Major: Engineering Physics
|
|---|