Uncertainty Aware Deep Learning for Particle Accelerators
This work addresses reliability issues in particle accelerator operations, providing uncertainty-aware models for specific accelerators, but it is incremental as it applies existing uncertainty methods to new domains.
The paper tackled the problem of deep learning models making inaccurate predictions on out-of-distribution inputs in particle accelerators, by implementing distance-aware uncertainty estimation using Deep Gaussian Process Approximation methods, resulting in improved detection of errant beams and surrogate modeling with quantified confidence.
Standard deep learning models for classification and regression applications are ideal for capturing complex system dynamics. However, their predictions can be arbitrarily inaccurate when the input samples are not similar to the training data. Implementation of distance aware uncertainty estimation can be used to detect these scenarios and provide a level of confidence associated with their predictions. In this paper, we present results from using Deep Gaussian Process Approximation (DGPA) methods for errant beam prediction at Spallation Neutron Source (SNS) accelerator (classification) and we provide an uncertainty aware surrogate model for the Fermi National Accelerator Lab (FNAL) Booster Accelerator Complex (regression).