PIPPIN: Generating variable length full events from partons

TL;DR

A novel approach for directly generating full events at detector-level from parton-level information, leveraging cutting-edge machine learning techniques, providing a promising direction for enhanced precision in fast detector simulation.

Abstract

This paper presents a novel approach for directly generating full events at detector-level from parton-level information, leveraging cutting-edge machine learning techniques. To address the challenge of multiplicity variations between parton and reconstructed object spaces, we employ transformers, score-based models and normalizing flows. Our method tackles the inherent complexities of the stochastic transition between these two spaces and achieves remarkably accurate results. The combination of innovative techniques and the achieved accuracy demonstrates the potential of our approach in advancing the field and opens avenues for further exploration. This research contributes to the ongoing efforts in high-energy physics and generative modelling, providing a promising direction for enhanced precision in fast detector simulation.

Figures (22)

• Figure 1: Diagram of the global architecture and the training processes of the PIPPIN model. It is made of two Transformer Encoders, which encode the partons, a Multiplicity Predictor, which predicts the number of reconstructed objects, and a PIP-Droid Generator, which conditionally generate these reconstructed objects.
• Figure 2: The Transformer Encoder architecture (left), which encodes the parton point cloud and the PIP-Droid Generator architecture (right), which denoises random tokens with type encoding into reconstructed objects, conditioned on the parton point cloud.
• Figure 3: The Multiplicity Predictor architecture. The partons presence is predicted and residually added to the encoded parton point cloud. Then, a global representation of the new encoded parton point cloud is learnt and conditionally used to sample the output multiplicity.
• Figure 4: Marginal distributions of the learnt reco-level multiplicities. Left: The multiplicity of the leptons in the reconstructed objects. Right: The multiplicity of the jets in the reconstructed objects. The grey area corresponds to the original MC simulation and the orange line to the output of the PIPPIN model. The bottom plots show the ratios of the histograms with respect to MC and the uncertainties as shaded areas.
• Figure 5: 2D marginal distributions of the learnt reco-level multiplicities. Left: The multiplicity of the leptons in the reconstructed objects. Right: The multiplicity of the jets in the reconstructed objects. The $x$-axis corresponds to the original MC simulation and the $y$-axis to the associated output of the PIPPIN model.