A University of Hawaiʻi at Mānoa student-led team developed an algorithm to determine direction in complex two-dimensional data. The research was published Feb. 6 in AIP Advances as a featured article.
The algorithm helps scientists locate the source of neutrinos, which can reveal information about nuclear reactors, the sun, and cosmic events. This method provides a mathematical foundation for extracting direction from noisy, real-world data, with applications extending to astronomy, medical imaging, and weather mapping.
Jeffrey G. Yepez, a physics undergraduate, led the algorithm development. “This approach gives researchers a clearer mathematical foundation for extracting direction from noisy, real-world data,” Yepez said. “It is a tool that scales with technological improvements in detectors, computing power and data volume.”
The project began with simulated neutrino data to locate nuclear reactors. Further studies are underway to expand the method’s use.
The algorithm uses the Frobenius norm as a “distance formula” to compare rotated reference datasets with measured data. By identifying the rotation that produces the smallest difference, the algorithm reveals the most likely direction of a signal. Simulations show the method works well with high-resolution data and large datasets.
The team was guided by Professor John G. Learned and mentored by UH alumnus Viacheslav Li. Funding came from the Consortium for Monitoring, Technology and Verification.
Other authors on the paper include Jackson D. Seligman, Max A. A. Dornfest, and Brian C. Crow. The Department of Physics and Astronomy is part of UH Mānoa’s College of Natural Sciences.
