How to Trust Learned Loop Amplitudes

TL;DR

This paper presents a machine learning approach to create reliable amplitude surrogates for high-precision LHC predictions, ensuring accurate uncertainty estimates and targeted improvements across phase space regions.

Contribution

It introduces a method for calibrated uncertainty estimation and targeted refinement of machine-learned amplitude surrogates in particle physics.

Findings

Surrogates provide calibrated uncertainties even with non-Gaussian systematics.

Less accurate phase space regions can be identified effectively.

Precision in these regions can be improved reliably.

Abstract

Higher-order theory predictions are crucial for the precision LHC program, but the time-consuming amplitude evaluation challenges the corresponding Monte-Carlo simulations. Machine-learned amplitude surrogates can resolve this problem, if we can guarantee their precision over the entire phase space. First, we show that our surrogates provide a calibrated learned uncertainty, even for non-Gaussian systematics; second, we describe how less accurate phase space regions can be identified; third, we demonstrate how the precision in these regions can be improved reliably.