Advances in developing deep neural networks for finding primary vertices in proton-proton collisions at the LHC

TL;DR

This paper presents a new deep neural network approach, based on UNet architecture, for identifying primary vertices in proton-proton collision data at the LHC, outperforming previous methods in efficiency and false positive rates.

Contribution

Developed an end-to-end tracks-to-histogram deep neural network that improves primary vertex detection accuracy over traditional KDE-based methods and adapts well across different LHC experiments.

Findings

DNN achieves lower false positive rates at high efficiency.

Model performance minimally degrades with FP16 quantization.

Algorithms validated against standard vertex finders, showing comparable or better efficiency.

Abstract

We are studying the use of deep neural networks (DNNs) to identify and locate primary vertices (PVs) in proton-proton collisions at the LHC. Earlier work focused on finding primary vertices in simulated LHCb data using a hybrid approach that started with kernel density estimators (KDEs) derived heuristically from the ensemble of charged track parameters and predicted "target histogram" proxies, from which the actual PV positions are extracted. We have recently demonstrated that using a UNet architecture performs indistinguishably from a "flat" convolutional neural network model. We have developed an "end-to-end" tracks-to-hist DNN that predicts target histograms directly from track parameters using simulated LHCb data that provides better performance (a lower false positive rate for the same high efficiency) than the best KDE-to-hists model studied. This DNN also provides better efficiency than the default heuristic algorithm for the same low false positive rate. "Quantization" of this model, using FP16 rather than FP32 arithmetic, degrades its performance minimally. Reducing the number of UNet channels degrades performance more substantially. We have demonstrated that the KDE-to-hists algorithm developed for LHCb data can be adapted to ATLAS and ACTS data using two variations of the UNet architecture. Within ATLAS/ACTS, these algorithms have been validated against the standard vertex finder algorithm. Both variations produce PV-finding efficiencies similar to that of the standard algorithm and vertex-vertex separation resolutions that are significantly better.