Machine Learning Extraction of Tensor Polarization in Spin-1 Observables
ORAL
Abstract
Tensor polarization enhancement using radio frequency manipulation of the
spin-1 line shape can improve the figure of merit of specialized tensor polariza-
tion observables in scattering experiments, improving the overall figure of merit.
This improvement is optimized only if careful manipulation and measurement
are synchronized to near-real time so that the RF manipulation can be applied
after each NMR sweep. This presentation discusses the measurement scheme
designed to reduce uncertainties in the tensor polarization while maximizing the
enhancement as well as a novel machine learning approach designed to improve
inference speed, accuracy, and precision.
spin-1 line shape can improve the figure of merit of specialized tensor polariza-
tion observables in scattering experiments, improving the overall figure of merit.
This improvement is optimized only if careful manipulation and measurement
are synchronized to near-real time so that the RF manipulation can be applied
after each NMR sweep. This presentation discusses the measurement scheme
designed to reduce uncertainties in the tensor polarization while maximizing the
enhancement as well as a novel machine learning approach designed to improve
inference speed, accuracy, and precision.
–
Presenters
-
Devin Allen Seay
University of Virginia
Authors
-
Devin Allen Seay
University of Virginia
-
Dustin Keller
University of Virginia