Distance-Weighted Graph Neural Networks on FPGAs for Real-Time Particle Reconstruction in High Energy Physics

08/08/2020
∙
by   Yutaro Iiyama, et al.
∙
0
∙

Graph neural networks have been shown to achieve excellent performance for several crucial tasks in particle physics, such as charged particle tracking, jet tagging, and clustering. An important domain for the application of these networks is the FGPA-based first layer of real-time data filtering at the CERN Large Hadron Collider, which has strict latency and resource constraints. We discuss how to design distance-weighted graph networks that can be executed with a latency of less than 1Ξs on an FPGA. To do so, we consider a representative task associated to particle reconstruction and identification in a next-generation calorimeter operating at a particle collider. We use a graph network architecture developed for such purposes, and apply additional simplifications to match the computing constraints of Level-1 trigger systems, including weight quantization. Using the 𝚑𝚕𝚜4𝚖𝚕 library, we convert the compressed models into firmware to be implemented on an FPGA. Performance of the synthesized models is presented both in terms of inference accuracy and resource usage.

READ FULL TEXT

page 1

page 2

page 3

page 4

∙ 04/04/2022

End-to-end multi-particle reconstruction in high occupancy imaging calorimeters with graph neural networks

We present an end-to-end reconstruction algorithm to build particle cand...
∙ 04/16/2018

Fast inference of deep neural networks in FPGAs for particle physics

Recent results at the Large Hadron Collider (LHC) have pointed to enhanc...
∙ 02/21/2019

Learning representations of irregular particle-detector geometry with distance-weighted graph networks

We explore the use of graph networks to deal with irregular-geometry det...
∙ 03/15/2022

Innovations in trigger and data acquisition systems for next-generation physics facilities

Data-intensive physics facilities are increasingly reliant on heterogene...
∙ 07/05/2021

Particle Convolution for High Energy Physics

We introduce the Particle Convolution Network (PCN), a new type of equiv...
∙ 12/15/2021

Domain-informed neural networks for interaction localization within astroparticle experiments

This work proposes a domain-informed neural network architecture for exp...
∙ 06/08/2020

Lorentz Group Equivariant Neural Network for Particle Physics

We present a neural network architecture that is fully equivariant with ...