Parameter Estimation using Neural Networks in the Presence of Detector Effects

TL;DR

This paper introduces SRGN, a two-level neural network reweighting method that reduces the need for multiple detector-simulated datasets in high energy physics parameter estimation, enabling more efficient and flexible fits.

Contribution

The paper presents SRGN, a novel two-level fitting approach that requires only one detector-simulated dataset, simplifying the process of parameter estimation in high energy physics.

Findings

SRGN effectively estimates parameters in various simulated scenarios.

The method reduces computational costs by minimizing the need for multiple detector simulations.

SRGN demonstrates comparable accuracy to traditional methods in top quark mass extraction.

Abstract

Histogram-based template fits are the main technique used for estimating parameters of high energy physics Monte Carlo generators. Parametrized neural network reweighting can be used to extend this fitting procedure to many dimensions and does not require binning. If the fit is to be performed using reconstructed data, then expensive detector simulations must be used for training the neural networks. We introduce a new two-level fitting approach that only requires one dataset with detector simulation and then a set of additional generation-level datasets without detector effects included. This Simulation-level fit based on Reweighting Generator-level events with Neural networks (SRGN) is demonstrated using simulated datasets for a variety of examples including a simple Gaussian random variable, parton shower tuning, and the top quark mass extraction.