First studies with quantitative machine learning in LHCb

First studies with quantitative machine learning in LHCb

Labeling the performance of the algorithm ( القوة_tag force-tagging) as a function of the tangential momentum p_T of the planes. Credit: University of Liverpool

The LHCb experiment at CERN recently announced the first proton-proton collision at world record energy with its all-new detector designed to handle the most demanding data-taking conditions.

The Data Processing and Analysis (DPA) project, led by University of Liverpool research physicist Eduardo Rodrigues, is a major overhaul of the offline analysis framework to allow full exploitation of the significant increase in data flow from the upgraded LHCb detector.

In a paper published in Journal of High Energy Physics, the DPA team demonstrated for the first time the successful use of quantum machine learning (QML) techniques to determine the charge of b-quark jets released into the LHC. This work is part of the research and development after the new data-taking period that has just begun, in the medium and long term.

Taking advantage of machine learning techniques are ubiquitous in analysis in LHCb. Due to the rapid progress of quantum computers and Quantum Technologiesit is only natural that we begin to investigate if and how quantum algorithms can be implemented on such new devices, and whether the LHCb Particle physics Use cases can benefit from new technology and a paradigm that is quantum computing.

So far, QML techniques have been applied primarily in particle physics to solve event classification problems and particle trajectory reconstruction, but the team applied them for the first time in the task of determining the hadron jet charge.

A ‘Quantum Machine Learning for B-jet Charge Determination’ study was conducted based on a sample simulation of launched b-quark aircraft. The performance of the so-called variable quantum classifier, based on two different quantitative circuits, was compared with the performance obtained using Deep Neural Network (DNN), which is a modern, classic (i.e. non-quantitative) type and a powerful type of artificial intelligence. algorithm. Performance is evaluated on a quantum simulator as the quantum devices available today are still in their early stage, although tests on real devices are currently in development.

The results, which were compared with those obtained using the classical DNN, showed that the performance of the DNN was slightly better than the QML algorithms, and the difference was small.

The paper shows that the QML method reaches optimal performance with fewer events, which helps reduce resource usage which will become a major point in the LHCb with the amount of data collected in the coming years. However, when using a large number of features, DNN performs better than QML algorithms. Improvements are expected when more high-performance quantum devices become available.

Studies conducted in collaboration with experts have shown that Quantum Algorithms It can allow the study of correlations between features. This can give the possibility of extracting information on the correlations of the jet components which will ultimately increase the performance of the jet flavor identification.

Dr. Eduardo Rodriguez says that “this paper demonstrated, for the first time, that QML can be successfully used to analyze LHCb data.” The exploitation of QML in particle physics experiments is still in its infancy. As physicists gain experience in quantum computing, drastic improvements in hardware and computing technology are expected due to global interest and investment in quantum computing.

“This work, which is part of the research and development activities of the LHCb Data Processing & Analysis (DPA) project, provided insight into QML. Exciting (first) results open new avenues for classification problems in particle physics experiments.”

Advances in algorithms make small and noisy quantum computers viable

more information:
Alessio Gianelle et al, Quantum machine learning for b-jet charge determination, Journal of High Energy Physics (2022). DOI: 10.1007/JHEP08 (2022) 014

Introduction of
University of Liverpool

the quote: First studies with quantitative machine learning at LHCb (2022, August 4) Retrieved August 5, 2022 from

This document is subject to copyright. Notwithstanding any fair dealing for the purpose of private study or research, no part may be reproduced without written permission. The content is provided for informational purposes only.