Search

Profile likelihoods, for instance, describing global SMEFT analyses at the LHC are numerically expensive to construct and evaluate. Especially profiled likelihoods are notoriously unstable and noisy. We show how modern numerical tools, similar to neural importance sampling, lead to a huge numerical improvement and allow us to evaluate the complete SFitter SMEFT likelihood in five hours on a single GPU., Comment: 24 pages, 9 figures

Differentiable programming opens exciting new avenues in particle physics, also affecting future event generators. These new techniques boost the performance of current and planned MadGraph implementations. Combining phase-space mappings with a set of very small learnable flow elements, MadNIS-Lite, can improve the sampling efficiency while being physically interpretable. This defines a third sampling strategy, complementing VEGAS and the full MadNIS., Comment: 16 pages, 6 figures

Recent innovations from machine learning allow for data unfolding, without binning and including correlations across many dimensions. We describe a set of known, upgraded, and new methods for ML-based unfolding. The performance of these approaches are evaluated on the same two datasets. We find that all techniques are capable of accurately reproducing the particle-level spectra across complex observables. Given that these approaches are conceptually diverse, they offer an exciting toolkit for a new class of measurements that can probe the Standard Model with an unprecedented level of detail and may enable sensitivity to new phenomena.

In pursuit of precise and fast theory predictions for the LHC, we present an implementation of the MadNIS method in the MadGraph event generator. A series of improvements in MadNIS further enhance its efficiency and speed. We validate this implementation for realistic partonic processes and find significant gains from using modern machine learning in event generators., Comment: 15 pages, 6 figures, 2 tables; v3: updates incl. referee requests

The matrix element method is the LHC inference method of choice for limited statistics. We present a dedicated machine learning framework, based on efficient phase-space integration, a learned acceptance and transfer function. It is based on a choice of INN and diffusion networks, and a transformer to solve jet combinatorics. We showcase this setup for the CP-phase of the top Yukawa coupling in associated Higgs and single-top production., Comment: 26 pages, 12 figures, v2: update references, v3: include evaluation on Herwig

Optimal kinematic observables are often defined in specific frames and then approximated at the reconstruction level. We show how multi-dimensional unfolding methods allow us to reconstruct these observables in their proper rest frame and in a probabilistically faithful way. We illustrate our approach with a measurement of a CP-phase in the top Yukawa coupling. Our method makes use of key advantages of generative unfolding, but as a constructed observable it fits into standard LHC analysis frameworks., Comment: 25 pages, 7 figures

Well-trained classifiers and their complete weight distributions provide us with a well-motivated and practicable method to test generative networks in particle physics. We illustrate their benefits for distribution-shifted jets, calorimeter showers, and reconstruction-level events. In all cases, the classifier weights make for a powerful test of the generative network, identify potential problems in the density estimation, relate them to the underlying physics, and tie in with a comprehensive precision and uncertainty treatment for generative networks., Comment: 32 pages, 19 figures

Theory predictions for the LHC require precise numerical phase-space integration and generation of unweighted events. We combine machine-learned multi-channel weights with a normalizing flow for importance sampling, to improve classical methods for numerical integration. We develop an efficient bi-directional setup based on an invertible network, combining online and buffered training for potentially expensive integrands. We illustrate our method for the Drell-Yan process with an additional narrow resonance., Comment: 33 pages, 15 figures, minor fixes to v1

Modern machine learning is transforming particle physics fast, bullying its way into our numerical tool box. For young researchers it is crucial to stay on top of this development, which means applying cutting-edge methods and tools to the full range of LHC physics problems. These lecture notes lead students with basic knowledge of particle physics and significant enthusiasm for machine learning to relevant applications. They start with an LHC-specific motivation and a non-standard introduction to neural networks and then cover classification, unsupervised classification, generative networks, and inverse problems. Two themes defining much of the discussion are well-defined loss functions and uncertainty-aware networks. As part of the applications, the notes include some aspects of theoretical LHC physics. All examples are chosen from particle physics publications of the last few years., Comment: Further expanded v3, we very much appreciate feedback

The matrix element method is widely considered the ultimate LHC inference tool for small event numbers. We show how a combination of two conditional generative neural networks encodes the QCD radiation and detector effects without any simplifying assumptions, while keeping the computation of likelihoods for individual events numerically efficient. We illustrate our approach for the CP-violating phase of the top Yukawa coupling in associated Higgs and single-top production. Currently, the limiting factor for the precision of our approach is jet combinatorics., Comment: 25 pages, 13 figures

First-principle simulations are at the heart of the high-energy physics research program. They link the vast data output of multi-purpose detectors with fundamental theory predictions and interpretation. This review illustrates a wide range of applications of modern machine learning to event generation and simulation-based inference, including conceptional developments driven by the specific requirements of particle physics. New ideas and tools developed at the interface of particle physics and machine learning will improve the speed and precision of forward simulations, handle the complexity of collision data, and enhance inference as an inverse simulation problem., Comment: Review article based on a Snowmass 2021 contribution

Generative networks are opening new avenues in fast event generation for the LHC. We show how generative flow networks can reach percent-level precision for kinematic distributions, how they can be trained jointly with a discriminator, and how this discriminator improves the generation. Our joint training relies on a novel coupling of the two networks which does not require a Nash equilibrium. We then estimate the generation uncertainties through a Bayesian network setup and through conditional data augmentation, while the discriminator ensures that there are no systematic inconsistencies compared to the training data., Comment: 28 pages, 14 figures

QCD splittings are among the most fundamental theory concepts at the LHC. We show how they can be studied systematically with the help of invertible neural networks. These networks work with sub-jet information to extract fundamental parameters from jet samples. Our approach expands the LEP measurements of QCD Casimirs to a systematic test of QCD properties based on low-level jet observables. Starting with an toy example we study the effect of the full shower, hadronization, and detector effects in detail., Comment: 25 pages, 11 figures

Autoencoder networks, trained only on QCD jets, can be used to search for anomalies in jet-substructure. We show how, based either on images or on 4-vectors, they identify jets from decays of arbitrary heavy resonances. To control the backgrounds and the underlying systematics we can de-correlate the jet mass using an adversarial network. Such an adversarial autoencoder allows for a general and at the same time easily controllable search for new physics. Ideally, it can be trained and applied to data in the same phase space region, allowing us to efficiently search for new physics using un-supervised learning., Comment: 11 figures, added references

The inclusion of the global effects of semiconductor doping poses a unique challenge for first-principles simulations, because the typically low concentration of dopants renders an explicit treatment intractable. In systems which do not display long-range band bending, a satisfactory remedy is offered by the use of "pseudoatoms", with a fractional nuclear charge matching the bulk doping concentration. However, this alone is not always sufficient for charged surfaces, where the width of the space-charge region (SCR) often exceeds realistic supercell dimensions. One generalization of the pseudoatom approach which overcomes this difficulty relies on the introduction of an artificially high doping level within a slab calculation, in conjunction with a multi-scale electrostatic energy correction. Here, we present an alternative technique that naturally extends the pseudoatom approach while bypassing the need for calculations with an unrealistically high doping level. It is based on the introduction of excess charge, mimicking free charge carriers from the SCR, along with a fixed sheet of counter-charge mimicking the SCR-related field. Self-consistency is obtained by imposing charge conservation and Fermi level equilibration between the bulk, treated semi-classically, and the electronic states of the slab/surface, which are treated quantum-mechanically. The method, which we call CREST - the Charge-Reservoir Electrostatic Sheet Technique - can be used with standard electronic structure codes. We validate the approach using a simple tight-binding model, which allows for comparison of its results with calculations encompassing the full SCR explicitly. We then employ it with density functional theory, where it is used to obtain insights into the electronic structures of the "clean-cleaved" Si(111) surface and its buckled (2x1) reconstruction, at various doping densities., Comment: 28 pages, 8 figures

The absorption and emission spectra of most luminescent, pi-conjugated, organic molecules are the mirror image of each other. In some cases, however, this symmetry is severely broken. In the present work, the asymmetry between the absorption and fluorescence spectra in molecular systems consisting of para-linked phenyl rings is studied. The vibronic structure of the emission and absorption bands is calculated from ab-initio quantum chemical methods and a subsequent, rigorous Franck-Condon treatment. Good agreement with experiment is achieved. A clear relation can be established between the strongly anharmonic double-well potential for the phenylene ring librations around the long molecular axis and the observed deviation from the mirror image symmetry. Consequences for related compounds and temperature dependent optical measurements are also discussed., Comment: 12 pages, 13 Figures

We study the dynamics of the `batch' minority game with market-impact correction using generating functional techniques to carry out the quenched disorder average. We find that the assumption of weak long-term memory, which one usually makes in order to calculate ergodic stationary states, breaks down when the persistent autocorrelation becomes larger than c_c=0.772... We show that this condition, remarkably, coincides with the AT-line found in an earlier static calculation. This result suggests a new scenario for ergodicity breaking in disordered systems., Comment: 6 pages, 3 figures, final version. To appear in J.Phys.A

We solve the dynamics of the on-line minority game, with general types of decision noise, using generating functional techniques a la De Dominicis and the temporal regularization procedure of Bedeaux et al. The result is a macroscopic dynamical theory in the form of closed equations for correlation- and response functions defined via an effective continuous-time single-trader process, which are exact in both the ergodic and in the non-ergodic regime of the minority game. Our solution also explains why, although one cannot formally truncate the Kramers-Moyal expansion of the process after the Fokker-Planck term, upon doing so one still finds the correct solution, that the previously proposed diffusion matrices for the Fokker-Planck term are incomplete, and how previously proposed approximations of the market volatility can be traced back to ergodicity assumptions., Comment: 25 pages LaTeX, no figures

We study the dynamics of a version of the batch minority game, with random external information and with different types of inhomogeneous decision noise (additive and multiplicative), using generating functional techniques \`{a} la De Dominicis. The control parameters in this model are the ratio $\alpha=p/N$ of the number $p$ of possible values for the external information over the number $N$ of trading agents, and the statistical properties of the agents' decision noise parameters. The presence of decision noise is found to have the general effect of damping macroscopic oscillations, which explains why in certain parameter regions it can effectively reduce the market volatility, as observed in earlier studies. In the limit $N\to\infty$ we (i) solve the first few time steps of the dynamics (for any $\alpha$), (ii) calculate the location $\alpha_c$ of the phase transition (signaling the onset of anomalous response), and (iii) solve the statics for $\alpha>\alpha_c$. We find that $\alpha_c$ is not sensitive to additive decision noise, but we arrive at non-trivial phase diagrams in the case of multiplicative noise. Our theoretical results find excellent confirmation in numerical simulations., Comment: 14 pages LaTeX, 8 postscript figures, in PRE style (minor textual changes in revision)

We study the dynamics of supervised on-line learning of realizable tasks in feed-forward neural networks. We focus on the regime where the number of examples used for training is proportional to the number of input channels N. Using generating function techniques from spin glass theory, we are able to average over the composition of the training set and transform the problem for N to infinity to an effective single pattern system, described completely by the student autocovariance, the student-teacher overlap and the student response function, with exact closed equations. Our method applies to arbitrary learning rules, i.e. not necessarily of a gradient-descent type. The resulting exact macroscopic dynamical equations can be integrated without finite-size effects up to any degree of accuracy, but their main value is in providing an exact and simple starting point for analytical approximation schemes. Finally, we show how, in the region of absent anomalous response and using the hypothesis that (as in detailed balance systems) the short-time part of the various operators can be transformed away, one can describe the stationary state of the network succesfully by a set of coupled equations involving only four scalar order parameters., Comment: 16 pages, 5 figures, uses iopart class

We study the dynamics of the batch minority game, with random external information, using generating functional techniques a la De Dominicis. The relevant control parameter in this model is the ratio $\alpha=p/N$ of the number $p$ of possible values for the external information over the number $N$ of trading agents. In the limit $N\to\infty$ we calculate the location $\alpha_c$ of the phase transition (signaling the onset of anomalous response), and solve the statics for $\alpha>\alpha_c$ exactly. The temporal correlations in global market fluctuations turn out not to decay to zero for infinitely widely separated times. For $\alpha<\alpha_c$ the stationary state is shown to be non-unique. For $\alpha\to 0$ we analyse our equations in leading order in $\alpha$, and find asymptotic solutions with diverging volatility $\sigma=\order(\alpha^{-{1/2}})$ (as regularly observed in simulations), but also asymptotic solutions with vanishing volatility $\sigma=\order(\alpha^{{1/2}})$. The former, however, are shown to emerge only if the agents' initial strategy valuations are below a specific critical value., Comment: 15 pages, 6 figures, uses Revtex. Replaced an old version of volatility graph that. Rephrased and updated some references

We study the dynamics of on-line learning in large perceptrons, for the case of training sets with a structural bias of the input vectors, by deriving exact and closed macroscopic dynamical laws using non-equilibrium statistical mechanical tools. In sharp contrast to the more conventional theories developed for homogeneously distributed or only weakly biased data, these laws are found to describe a non-trivial and persistently non-deterministic macroscopic evolution, and a generalisation error which retains both stochastic and sample-to-sample fluctuations, even for infinitely large networks. Furthermore, for the standard error-correcting microscopic algorithms (such as the perceptron learning rule) one obtains learning curves with distinct bias-induced phases. Our theoretical predictions find excellent confirmation in numerical simulations., Comment: 20 pages

We study the magnetic properties of nanoscale magnetic films with large perpendicular anisotropy comparing polarization microscopy measurements on Co_28Pt_72 alloy samples based on the magneto-optical Kerr effect with Monte Carlo simulations of a corresponding micromagnetic model. We focus on the understanding of the dynamics especially the temperature and field dependence of the magnetisation reversal process. The experimental and simulational results for hysteresis, the reversal mechanism, domain configurations during the reversal, and the time dependence of the magnetisation are in very good qualitative agreement. The results for the field and temperature dependence of the domain wall velocity suggest that for thin films the hysteresis can be described as a depinning transition of the domain walls rounded by thermal activation for finite temperatures., Comment: 7 pages Latex, Postscript figures included, accepted for publication in Phys.Rev.B, also availible at: http://www.thp.Uni-Duisburg.DE/Publikationen/Publist_Us_R.html