Background Developing deep learning models for radiology requires large data sets and substantial computational resources. Data set size limitations can be further exacerbated by distribution shifts, such as rapid changes in patient populations and standard of care during the COVID-19 pandemic. A common partial mitigation is transfer learning by pretraining a "generic network" on a large nonmedical data set and then fine-tuning on a task-specific radiology data set.
View Article and Find Full Text PDFWe report the measurement of the beam-vector and tensor asymmetries A_{ed}^{V} and A_{d}^{T} in quasielastic (e[over →],e^{'}p) electrodisintegration of the deuteron at the MIT-Bates Linear Accelerator Center up to missing momentum of 500 MeV/c. Data were collected simultaneously over a momentum transfer range 0.1 View Article and Find Full Text PDF
We report a precision measurement of the deuteron tensor analyzing powers T(20) and T(21) at the MIT-Bates Linear Accelerator Center. Data were collected simultaneously over a momentum transfer range Q=2.15-4.
View Article and Find Full Text PDFWe report new measurements of the neutron charge form factor at low momentum transfer using quasielastic electrodisintegration of the deuteron. Longitudinally polarized electrons at an energy of 850 MeV were scattered from an isotopically pure, highly polarized deuterium gas target. The scattered electrons and coincident neutrons were measured by the Bates Large Acceptance Spectrometer Toroid (BLAST) detector.
View Article and Find Full Text PDFWe report the first precision measurement of the proton electric to magnetic form factor ratio from spin-dependent elastic scattering of longitudinally polarized electrons from a polarized hydrogen internal gas target. The measurement was performed at the MIT-Bates South Hall Ring over a range of four-momentum transfer squared Q2 from 0.15 to 0.
View Article and Find Full Text PDF