MadNIS at NLO

TL;DR

This paper introduces MadNIS, a method combining neural importance sampling and amplitude surrogates to significantly speed up and improve the accuracy of next-to-leading order (NLO) calculations in particle physics.

Contribution

It presents a novel approach that integrates neural surrogates with traditional NLO techniques, enabling faster and more reliable computations across phase space.

Findings

Achieved significant speed-ups in NLO calculations for electron-positron scattering.

Demonstrated variance reduction in the integration process.

Validated the method for three- and four-jet processes.

Abstract

We combine fast amplitude surrogates with neural importance sampling to accelerate NLO calculations. For virtual corrections, a learned ratio to the Born matrix element with calibrated uncertainties guarantees reliable precision across phase space. For real emission, we stick to the standard FKS subtraction and train sector-conditioned surrogates of the regularized integrands away from divergences. MadNIS then uses multi-channel mappings and FKS sectors as conditions. We validate our approach for electron-positron scattering to three and four jets and find significant speed-ups and variance reduction in the integration.