nFacet Analysis

This is a repository with notes about various aspects of the nFacet project.

Contents:

• Gamma Calibration
  • Introduction and loading data
  • Channel fitting
• Neural network for dose measurement
  • Introduction and motivation
  • Training inputs
  • Training targets
  • Choice of training dataset
  • Neural network architecture
  • Analysis of error
  • Fluence binning scheme
  • Cubes only model, 500 epochs, 200 samples per dataset
  • Cubes and profiles model, 200 samples per dataset, learning rate 0.001
  • Cubes model, learning rate 0.001, 200 samples per dataset, trained for 2000 epochs
  • Cubes and profiles model, learning rate 0.001, 200 samples per dataset, trained for 2000 epochs
  • Current results & future steps
• Comparisons between real and simulated data
• Simulated LiF screen study
  • Goals of the study