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

Abstract

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 1$\mu \mathrm{s}$ 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.

arXiv comment15 pages, 4 figures

References in corpus (13)

• Adam: A Method for Stochastic Optimization
• Array Programming with NumPy
• The NumPy array: a structure for efficient numerical computation
• Particle-flow reconstruction and global event description with the CMS detector
• Fast inference of deep neural networks in FPGAs for particle physics
• ParticleNet: Jet Tagging via Particle Clouds
• Jet reconstruction and performance using particle flow with the ATLAS Detector
• Learning representations of irregular particle-detector geometry with distance-weighted graph networks
• JEDI-net: a jet identification algorithm based on interaction networks
• Compressing deep neural networks on FPGAs to binary and ternary precision with HLS4ML
• Fast inference of Boosted Decision Trees in FPGAs for particle physics
• Interaction networks for the identification of boosted $H\to b\overline{b}$ decays
• Probing stop pair production at the LHC with graph neural networks

Cited by in corpus (20)

• Automatic heterogeneous quantization of deep neural networks for low-latency inference on the edge for particle detectors
• Fast convolutional neural networks on FPGAs with hls4ml
• Anomaly detection with Convolutional Graph Neural Networks
• Applications and Techniques for Fast Machine Learning in Science
• Autoencoders on FPGAs for real-time, unsupervised new physics detection at 40 MHz at the Large Hadron Collider
• MLPF: Efficient machine-learned particle-flow reconstruction using graph neural networks
• Charged particle tracking via edge-classifying interaction networks
• Graph Neural Networks for Charged Particle Tracking on FPGAs
• Nanosecond machine learning event classification with boosted decision trees in FPGA for high energy physics
• LL-GNN: Low Latency Graph Neural Networks on FPGAs for High Energy Physics
• A Detailed Study of Interpretability of Deep Neural Network based Top Taggers
• Nanosecond machine learning regression with deep boosted decision trees in FPGA for high energy physics
• Real-time Graph Building on FPGAs for Machine Learning Trigger Applications in Particle Physics
• Calorimetric Measurement of Multi-TeV Muons via Deep Regression
• Ultrafast jet classification on FPGAs for the HL-LHC
• Real-time Artificial Intelligence for Accelerator Control: A Study at the Fermilab Booster
• Accelerating Graph-based Tracking Tasks with Symbolic Regression
• BitHEP -- The Limits of Low-Precision ML in HEP
• Real-time graph neural networks on FPGAs for the Belle II electromagnetic calorimeter
• End-to-end multi-particle reconstruction in high occupancy imaging calorimeters with graph neural networks