Shared Data and Algorithms for Deep Learning in Fundamental Physics

TL;DR

This paper presents a unified Python package offering access to diverse fundamental physics datasets for machine learning, along with a versatile graph neural network approach that achieves near state-of-the-art performance across tasks.

Contribution

The authors introduce a unified data interface and reference models for fundamental physics datasets, facilitating cross-disciplinary machine learning and transfer learning research.

Findings

Achieves performance close to specialized methods on all datasets.

Provides a flexible graph neural network architecture adaptable to various physics problems.

Simplifies data handling and model implementation for fundamental physics machine learning.

Abstract

We introduce a Python package that provides simply and unified access to a collection of datasets from fundamental physics research - including particle physics, astroparticle physics, and hadron- and nuclear physics - for supervised machine learning studies. The datasets contain hadronic top quarks, cosmic-ray induced air showers, phase transitions in hadronic matter, and generator-level histories. While public datasets from multiple fundamental physics disciplines already exist, the common interface and provided reference models simplify future work on cross-disciplinary machine learning and transfer learning in fundamental physics. We discuss the design and structure and line out how additional datasets can be submitted for inclusion. As showcase application, we present a simple yet flexible graph-based neural network architecture that can easily be applied to a wide range of supervised learning tasks. We show that our approach reaches performance close to dedicated methods on all datasets. To simplify adaptation for various problems, we provide easy-to-follow instructions on how graph-based representations of data structures, relevant for fundamental physics, can be constructed and provide code implementations for several of them. Implementations are also provided for our proposed method and all reference algorithms.