Machine-learned particle flow as a foundation model for collider physics

2026-06-12

AI Summary

This work addresses the fragmented and modular nature of traditional particle collision event reconstruction pipelines, which hinder end-to-end modeling from detector data to physics analysis. The authors formulate reconstruction as a machine learning problem by leveraging the Machine Learning Particle Flow (MLPF) framework to extract a unified particle-level latent representation. This representation is then adapted to multiple downstream tasks—including jet flavor tagging, jet energy regression, and missing transverse momentum regression—using only a single linear layer per task. Experimental results demonstrate that the proposed approach achieves superior performance on tasks such as missing transverse momentum regression while drastically reducing model complexity, with only approximately 1/35 the number of parameters compared to baseline methods. This study provides the first evidence that latent representations learned during reconstruction can serve as a general-purpose foundation model for efficient and unified high-energy physics analysis.

Abstract

The workflow from particle collision to physics analysis passes through a series of reconstruction steps that are traditionally modular and disconnected, with no shared representation linking low-level detector data to high-level analysis tasks. We show that casting event reconstruction as a machine learning problem naturally produces such a shared representation. We repurpose a machine learning model trained for particle-flow reconstruction (MLPF) to perform three distinct analysis tasks: jet flavor identification, jet energy regression, and missing momentum regression. By appending the per-particle latent representations learned during reconstruction as additional input features, we substantially improve over baselines that use kinematic features alone. We further demonstrate that a single linear layer trained using only the latent representations achieves competitive performance against state-of-the-art baseline architectures, and outperforms the baseline for missing momentum regression with approximately 35 times fewer parameters. These results demonstrate that the latent representations learned during reconstruction encode essential physics information needed for downstream analysis, establishing MLPF as a foundation model and offering a concrete step toward an end-to-end pipeline from detector data to physics analysis.

Problem

• particle flow
• event reconstruction
• foundation model
• latent representation
• collider physics

Innovation

• foundation model
• particle flow
• latent representation
• end-to-end reconstruction
• machine learning for physics