arXiv subjects

Marian Stahl

Publications and source records attributed to Marian Stahl.

4 recordsLinked to original sources

arXiv2026

A converged architecture for processing 32 Tbps of physics data in real-time at the LHCb experiment

The LHCb detector at the Large Hadron Collider has been upgraded to acquire an unprecedented 32 Tbps of particle-collision data to provide new insights in the High Energy Physics domain. The data produced by the detector is filtered in real-time to select interesting collisions. As part of the upgrade, a pre-filtering stage has been removed leading to a factor 40 increase in data rate. To deal with the high throughput demands of LHCb real-time data processing, we present an off-the-shelf network architecture using zero-copy techniques in conjunction with an efficient, fully-GPU-based filter. Our converged architecture is able to process the full 32 Tbps of particle-collision data in real-time, the highest in any physics experiment to date. Our result extends the reach of the LHCb physics programme and sets a new standard for real-time data processing at particle physics experiments.

hep-ex↗

arXiv2021

Progress in developing a hybrid deep learning algorithm for identifying and locating primary vertices

The locations of proton-proton collision points in LHC experiments are called primary vertices (PVs). Preliminary results of a hybrid deep learning algorithm for identifying and locating these, targeting the Run 3 incarnation of LHCb, have been described at conferences in 2019 and 2020. In the past year we have made significant progress in a variety of related areas. Using two newer Kernel Density Estimators (KDEs) as input feature sets improves the fidelity of the models, as does using full LHCb simulation rather than the "toy Monte Carlo" originally (and still) used to develop models. We have also built a deep learning model to calculate the KDEs from track information. Connecting a tracks-to-KDE model to a KDE-to-hists model used to find PVs provides a proof-of-concept that a single deep learning model can use track information to find PVs with high efficiency and high fidelity. We have studied a variety of models systematically to understand how variations in their architectures affect performance. While the studies reported here are specific to the LHCb geometry and operating conditions, the results suggest that the same approach could be used by the ATLAS and CMS experiments.

hep-ex↗

arXiv2020

An updated hybrid deep learning algorithm for identifying and locating primary vertices

We present an improved hybrid algorithm for vertexing, that combines deep learning with conventional methods. Even though the algorithm is a generic approach to vertex finding, we focus here on it's application as an alternative Primary Vertex (PV) finding tool for the LHCb experiment. In the transition to Run 3 in 2021, LHCb will undergo a major luminosity upgrade, going from 1.1 to 5.6 expected visible PVs per event, and it will adopt a purely software trigger. We use a custom kernel to transform the sparse 3D space of hits and tracks into a dense 1D dataset, and then apply Deep Learning techniques to find PV locations using proxy distributions to encode the truth in training data. Last year we reported that training networks on our kernels using several Convolutional Neural Network layers yielded better than 90 % efficiency with no more than 0.2 False Positives (FPs) per event. Modifying several elements of the algorithm, we now achieve better than 94 % efficiency with a significantly lower FP rate. Where our studies to date have been made using toy Monte Carlo (MC), we began to study KDEs produced from complete LHCb Run 3 MC data, including full tracking in the vertex locator rather than proto-tracking.

physics.ins-det↗

arXiv2017

Machine learning and parallelism in the reconstruction of LHCb and its upgrade

After a highly successful first data taking period at the LHC, the LHCb experiment developed a new trigger strategy with a real-time reconstruction, alignment and calibration for Run II. This strategy relies on offline-like track reconstruction in the high level trigger, making a separate offline event reconstruction unnecessary. To enable such reconstruction, and additionally keeping up with a higher event rate due to the accelerator upgrade, the time used by the track reconstruction had to be decreased. Timing improvements have in parts been achieved by utilizing parallel computing techniques that will be described in this document by considering two example applications. Despite decreasing computing time, the reconstruction quality in terms of reconstruction efficiency and fake rate could be improved at several places. Two applications of fast machine learning techniques are highlighted, refining track candidate selection at the early stages of the reconstruction.

physics.ins-det↗