NOBLE is a neural operator framework introduced to capture experimental variability in biological neuron models.
It learns a mapping from neuron features to the voltage response, predicting distributions of neural dynamics accounting for experimental variability.
NOBLE offers models with dynamics consistent with observed responses and provides a $4200×$ speedup over numerical solvers.
It enables efficient generation of synthetic neurons with trial-to-trial variability, contributing to a better understanding of brain function and neuroAI applications.