Induced Generative Adversarial Particle Transformers

TL;DR

This paper introduces iGAPT, a novel particle transformer model that improves simulation of particle collisions by achieving linear time complexity and better capturing jet substructure, surpassing previous models like MPGAN.

Contribution

iGAPT integrates induced particle-attention blocks and global conditioning, offering a more efficient and accurate simulation method for high energy physics data.

Findings

iGAPT achieves linear time complexity, improving efficiency.

iGAPT surpasses MPGAN in multiple metrics.

iGAPT effectively captures complex jet substructure.

Abstract

In high energy physics (HEP), machine learning methods have emerged as an effective way to accurately simulate particle collisions at the Large Hadron Collider (LHC). The message-passing generative adversarial network (MPGAN) was the first model to simulate collisions as point, or ``particle'', clouds, with state-of-the-art results, but suffered from quadratic time complexity. Recently, generative adversarial particle transformers (GAPTs) were introduced to address this drawback; however, results did not surpass MPGAN. We introduce induced GAPT (iGAPT) which, by integrating ``induced particle-attention blocks'' and conditioning on global jet attributes, not only offers linear time complexity but is also able to capture intricate jet substructure, surpassing MPGAN in many metrics. Our experiments demonstrate the potential of iGAPT to simulate complex HEP data accurately and efficiently.