Stay Positive: Neural Refinement of Sample Weights

TL;DR

This paper presents a neural network-based weight refinement method for Monte Carlo simulations in particle physics, improving accuracy and simplicity over existing approaches, and introduces a compatible resampling protocol.

Contribution

It introduces a novel neural refinement approach that scales initial weights without full distribution modeling and a new resampling protocol preserving statistical properties.

Findings

Neural refinement matches or exceeds existing weight transformation accuracy.

The resampling protocol is simpler and maintains statistical uncertainties.

Method performs well on both synthetic and realistic data.

Abstract

Monte Carlo simulations are an essential tool in particle physics data analysis. Events are typically generated alongside weights that redistribute the cross section of the simulated process across the phase space. These weights can be negative, and several post-hoc methods have been developed to eliminate or mitigate the negative values. All of these methods share the common strategy of approximating the average weight as a function of phase space. We introduce an alternative approach, which, instead of reweighting to the average, refines the initial weights with a scaling transformation, utilizing a phase space-dependent factor. Since this new refinement method does not need to model the full weight distribution, it can be more accurate. High-dimensional and unbinned phase space is processed using neural networks for the refinement method. In addition to the refinement method, we introduce a new resampling protocol, which can be used in conjunction with any weight transformation to not only preserve the average weight but also the statistical uncertainties of the initial distribution. Using both realistic and synthetic examples, we show that the new neural refinement method is able to match or exceed the accuracy of similar weight transformations and that the new resampling protocol is simpler in implementation than previous methods while exhibiting equivalent statistical properties.