Particle Trajectory Representation Learning with Masked Point Modeling

2025-02-04

AI Summary

To address the challenge of learning sparse, high-dimensional, unlabeled 3D particle trajectory representations in liquid argon time projection chambers (LArTPCs), this paper proposes PoLAr-MAE, a self-supervised masked point modeling framework. Methodologically, it introduces a resolution-agnostic voxelization-based patching strategy, couples position and ionization energy infilling into a unified energy-aware masking-and-reconstruction task, and extends the PointMAE architecture to support multi-scale point cloud reconstruction. Contributions include: (i) achieving trajectory classification performance (F1 = 99.4%) and shower classification performance (F1 = 97.7%) on purely unsupervised LArTPC data—comparable to fully supervised methods; and (ii) releasing PILArNet-M, a large-scale benchmark dataset comprising one million LArTPC events and 5.2 billion annotated 3D points, thereby advancing cross-disciplinary research at the intersection of high-energy physics and self-supervised point cloud learning.

Technology Category

• Machine Learning: Unsupervised & Self-Supervised Learning
• Search and Optimization: Learning to Search
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Effective self-supervised learning (SSL) techniques have been key to unlocking large datasets for representation learning. While many promising methods have been developed using online corpora and captioned photographs, their application to scientific domains, where data encodes highly specialized knowledge, remains in its early stages. We present a self-supervised masked modeling framework for 3D particle trajectory analysis in Time Projection Chambers (TPCs). These detectors produce globally sparse (<1% occupancy) but locally dense point clouds, capturing meter-scale particle trajectories at millimeter resolution. Starting with PointMAE, this work proposes volumetric tokenization to group sparse ionization points into resolution-agnostic patches, as well as an auxiliary energy infilling task to improve trajectory semantics. This approach -- which we call Point-based Liquid Argon Masked Autoencoder (PoLAr-MAE) -- achieves 99.4% track and 97.7% shower classification F-scores, matching that of supervised baselines without any labeled data. While the model learns rich particle trajectory representations, it struggles with sub-token phenomena like overlapping or short-lived particle trajectories. To support further research, we release PILArNet-M -- the largest open LArTPC dataset (1M+ events, 5.2B labeled points) -- to advance SSL in high energy physics (HEP). Project site: https://youngsm.com/polarmae/

Problem

• Self-supervised learning for particle trajectories
• Masked modeling in Time Projection Chambers
• Improving trajectory semantics in sparse data

Innovation

• Self-supervised masked modeling framework
• Volumetric tokenization for sparse data
• Auxiliary energy infilling task