The WLCG has been essential to the success of the LHC’s scientific programme. It is used to store and analyse the data produced by the LHC experiments. Optimal usage of the grid’s resources is a major challenge: with the foreseen increase in the data produced by the LHC experiments, workflow optimisation — particularly for the data-placement strategy — becomes extremely important.

Simulating this complex and highly non-linear environment is very difficult; the complexity of the task goes beyond the capability of the computing hardware available today. Quantum computing could offer the possibility to address this. Our project, a collaboration with the Polytechnic Institute of Grenoble and the Polytechnic University of Bucharest, will develop quantum algorithms to optimise the storage distribution.

QSVMs are among the most promising machine-learning algorithms for quantum computers. Initial quantum implementations have already shown performances comparable to their classical counterparts. QSVMs are considered suitable algorithms for early adoption on noisy, near-term quantum-computing devices. Several initiatives are studying and optimising input data representation and training strategies.

We are testing IBM’s QSVM algorithm at the ATLAS experiment. Today, identifying ttH-production events relies on classical support vector machines, as well as another machine-learning technique known as ‘boosted decision trees’. Classically, these methods are used to improve event selection and background rejection by analysing 47 high-level characteristic features.

The analysis of LHC data for the detection of effects beyond the Standard Model requires increasing levels of precision. Various machine-learning techniques are now part of the standard analysis toolbox for high-energy physics. Deep-learning algorithms are increasingly demonstrating their usefulness in various areas of analysis, thanks to their suitability for exploring a much larger dimensional space.

This seems close to an ideal area of application for quantum computing, which offers a parameter space that is potentially enormous, as well as a correspondingly large level of computational parallelism. Moreover, the quasi-optimal Gibbs sampling features of quantum computers may enhance the training of the deep-learning networks.

The LHC at CERN is producing collisions at unprecedented collider energy. The hundreds of particles created during the collisions are recorded by large detectors composed of several sub-detectors. At the centre of these detectors there is usually a tracker detector, precisely recording the signal of the passage of charged particles through thin layers of active material. The trajectories of particles are bent by a magnetic field to allow the measurement of the particle momentum. There is an expected ten-fold increase in the number of tracks produced per bunch crossing after the high-luminosity upgrade of LHC.

Classical algorithms perform the reconstruction of the trajectory of charged particles; these make use of Kalman filter formalism. Although they are quite accurate, they scale worse than quadratically with the number of tracks. Several ways are explored to mitigate the increase in the computing needs, such as new detector layout, deep-learning and code parallelisation. Quantum computing has been shown to provide speed-ups for certain problems and different R&D initiatives are exploring how quantum tracking algorithms could exploit such capabilities. We are developing a quantum-based track-finding algorithm aimed at reducing the combinatorial background during the initial seeding stage for the Kalman filters. We are using the publicly available dataset designed for the Kaggle ‘TrackML’ challenge for this work.

Research in high-energy physics, as in many other scientific domains, makes extensive use of Monte Carlo calculations — both in theory and in experiments.

GANs are among the most interesting models in classical machine learning. GANs are an example of generative models (i.e. models that learn a hidden distribution from a training dataset) and can sample new synthetic data. We have been investigating their use as an alternative to Monte Carlo simulation, obtaining remarkable results. Much faster than standard Monte Carlo algorithms, GANs can generate realistic synthetic data, while retaining a high level of accuracy (see our fast simulation project). Quantum GANs could have more representational power than classical GANs, making them better able to learn more complex distributions from smaller training datasets. We are now training a quantum GAN to generate images of a few pixels. For this we are investigating two possible approaches:

• A hybrid schema with a quantum generator learning the target PDF, using either a classical network or a variational quantum circuit as a discriminator (variational quantum generator).
• A full quantum adversarial implementation (quGAN).

Monte Carlo methods make extensive use of random numbers. To ensure simulated data is unbiased, it is essential to understand the quality of the random number generators, the question of true randomness, and other related issues. Modern pseudo-random number generators exhibit optimal performance; we are studying the effect of replacing those with real quantum numbers in different Monte Carlo applications.