Simulation-Prior Independent Neural Unfolding Procedure

TL;DR

SPINUP is a neural unfolding method that independently reconstructs distributions from high-dimensional data at the LHC, reducing prior dependence and improving efficiency with neural importance sampling.

Contribution

The paper introduces SPINUP, a novel neural unfolding technique that is prior-independent and utilizes neural importance sampling for efficient high-dimensional data unfolding.

Findings

Successfully unfolded jet substructure observables.

Unfolded Higgs and single-top production to parton level.

Demonstrated reduced prior dependence in unfolding results.

Abstract

Machine learning allows unfolding high-dimensional spaces without binning at the LHC. The new SPINUP method extracts the unfolded distribution based on a neural network encoding the forward mapping, making it independent of the prior from the simulated training data. It is made efficient through neural importance sampling, and ensembling can be used to estimate the effect of information loss in the forward process. We showcase SPINUP for unfolding detector effects on jet substructure observables and for unfolding to parton level of associated Higgs and single-top production.