Phase Space Sampling and Inference from Weighted Events with Autoregressive Flows

TL;DR

This paper introduces autoregressive flows for efficient generation and inference of particle collider events, accommodating variable and negative event weights, demonstrated on top pair production at colliders.

Contribution

It presents a novel use of autoregressive flows with weighted likelihood for physical event generation, handling negative weights in collider data.

Findings

Effective generation of collider events with importance sampling weights.

Successful inference from events with negative weights.

Demonstrated applicability to top pair production at colliders.

Abstract

We explore the use of autoregressive flows, a type of generative model with tractable likelihood, as a means of efficient generation of physical particle collider events. The usual maximum likelihood loss function is supplemented by an event weight, allowing for inference from event samples with variable, and even negative event weights. To illustrate the efficacy of the model, we perform experiments with leading-order top pair production events at an electron collider with importance sampling weights, and with next-to-leading-order top pair production events at the LHC that involve negative weights.