Fast inference of Boosted Decision Trees in FPGAs for particle physics

Summers, S.; Guglielmo, Giuseppe Di; Duarte, J.; Harris, P.; Hoang, Duc; Jindariani, S.; Kreinar, Edward; Lončar, Vladimir; Ngadiuba, J.; Pierini, M.; Rankin, D.; Tran, N. V.; Wu, Z.

doi:10.1088/1748-0221/15/05/p05026

Cited by 58 publications

(41 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Development of ML models deployable to FPGA-based L1T systems is helped by tools for automatic network-to-circuit conversion such as hls4ml. Using hls4ml, several solutions for HEP-specific tasks (e.g., jet tagging) have been provided (Duarte et al, 2018;Coelho et al, 2020;Di Guglielmo et al, 2020;Summers et al, 2020), exploiting models with simpler architectures than what is shown here. This tool has been applied extensively for tasks in the HL-LHC upgrade of the CMS L1T system, including an autoencoder for anomaly detection, and DNNs for muon energy regression and identification, tau lepton identification, and vector boson fusion event classification (CMS Collaboration, 2020).…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

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

et al. 2021

View full text Add to dashboard Cite

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 one μs 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.

show abstract

Section: Related Workmentioning

confidence: 99%

“…To this end, its design prioritizes all-on-chip implementations of the most common network components. Its functionality has been demonstrated with dense neural networks (DNNs) ( Duarte et al, 2018 ), extended to also support BDTs ( Summers et al, 2020 ). Extensions to convolutional and recurrent neural networks are in development.…”

Section: Introductionmentioning

confidence: 99%

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

et al. 2021

View full text Add to dashboard Cite

show abstract

“…Work has also been done to accelerate the inference of deep neural networks with heterogeneous resources beyond GPUs, like field-programmable gate arrays (FPGAs) [49][50][51][52][53][54][55][56][57]. This work extends to GNN architectures [29,58].…”

Section: Inference Timingmentioning

confidence: 99%

Charged Particle Tracking via Edge-Classifying Interaction Networks

Dezoort

Thais

Duarte

et al. 2021

Comput Softw Big Sci

View full text Add to dashboard Cite

Recent work has demonstrated that geometric deep learning methods such as graph neural networks (GNNs) are well suited to address a variety of reconstruction problems in high-energy particle physics. In particular, particle tracking data are naturally represented as a graph by identifying silicon tracker hits as nodes and particle trajectories as edges, given a set of hypothesized edges, edge-classifying GNNs identify those corresponding to real particle trajectories. In this work, we adapt the physics-motivated interaction network (IN) GNN toward the problem of particle tracking in pileup conditions similar to those expected at the high-luminosity Large Hadron Collider. Assuming idealized hit filtering at various particle momenta thresholds, we demonstrate the IN’s excellent edge-classification accuracy and tracking efficiency through a suite of measurements at each stage of GNN-based tracking: graph construction, edge classification, and track building. The proposed IN architecture is substantially smaller than previously studied GNN tracking architectures; this is particularly promising as a reduction in size is critical for enabling GNN-based tracking in constrained computing environments. Furthermore, the IN may be represented as either a set of explicit matrix operations or a message passing GNN. Efforts are underway to accelerate each representation via heterogeneous computing resources towards both high-level and low-latency triggering applications.

show abstract

“…We now discuss some triggers constructed using the above BDT classification and quote the signal efficiency and background rates for each of them. In the Phase-II upgrade, the L1 trigger hardware will have FPGAs which will be able to handle small scale machine learning (ML) applications, like BDT classification [70,71]. We choose three points from the ROC curves, corresponding to 98%, 90% and 70% background rejections.…”

Section: Triggers Based On the Bdt Training Using Variables From Tracmentioning

confidence: 99%

Triggering long-lived particles in HL-LHC and the challenges in the first stage of the trigger system

Bhattacherjee

Mukherjee

Sengupta

et al. 2020

J. High Energ. Phys.

View full text Add to dashboard Cite

Triggering long-lived particles (LLPs) at the first stage of the trigger system is very crucial in LLP searches to ensure that we do not miss them at the very beginning. The future High Luminosity runs of the Large Hadron Collider will have increased number of pileup events per bunch crossing. There will be major upgrades in hardware, firmware and software sides, like tracking at level-1 (L1). The L1 trigger menu will also be modified to cope with pileup and maintain the sensitivity to physics processes. In our study we found that the usual level-1 triggers, mostly meant for triggering prompt particles, will not be very efficient for LLP searches in the 140 pileup environment of HL-LHC, thus pointing to the need to include dedicated L1 triggers in the menu for LLPs. We consider the decay of the LLP into jets and develop dedicated jet triggers using the track information at L1 to select LLP events. We show in our work that these triggers give promising results in identifying LLP events with moderate trigger rates.

show abstract

Fast inference of Boosted Decision Trees in FPGAs for particle physics

Cited by 58 publications

References 16 publications

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

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

Charged Particle Tracking via Edge-Classifying Interaction Networks

Triggering long-lived particles in HL-LHC and the challenges in the first stage of the trigger system

Contact Info

Product

Resources

About